Oxygen transfer during aerobic exercise in a varanid lizard Varanus mertensi is limited by the circulation

Metabolic rate, or more precisely, the rate of oxygen consumption(V̇_O2), is set by structural and functional parameters at each point in the oxygen pathway from lungs to the mitochondria in skeletal muscle. It is uniformly accepted that,in mammals, the `limitation' to maximal V̇_O2 is distributed over all components of the pathway, though under certain conditions some steps may provide more resistance than others(Hoppeler and Weibel,1998).

A number of studies have searched for the component(s) that limit oxygen transfer and hence aerobic metabolism in exercising varanid lizards, a group that is considered to have one of the highest aerobic scopes amongst reptiles(Gleeson et al., 1980;Mitchell et al., 1981a,b). Almost all of these studies have been restricted to one species, Varanus exanthematicus. It would appear that this species effectively ventilates its lungs to match the increase in metabolic rate achieved during sustained locomotion, at levels sufficient to achieve maximal rates of oxygen uptake(Hopkins et al., 1995; Wang et al., 1997). Indeed,ventilation has been shown to increase proportionally more than metabolic rate(Mitchell et al., 1981b; Wang et al., 1997), giving rise to an elevated alveolar partial pressure of oxygen(PA_O2)(Mitchell et al., 1981b). This suggests that aerobic activity in varanids is not limited by ventilation, as originally suggested in the axial constraint hypothesis proposed by Carrier(1987b). In contrast to the elevated PA_O2, arterial P_O2 (Pa_O2)remains somewhat constant (Mitchell et al., 1981b; Hopkins et al.,1995), implying that the observed alveolar—arterial P_O2 difference observed during exercise is potentially limited by pulmonary diffusion, the inability to increase diffusing capacity during exercise, ventilation/perfusion inequalities and/or cardiac shunting (Mitchell et al.,1981b; Hopkins et al.,1995).

It has been suggested that the greater factorial scope in V. exanthematicus, compared to other lizards such as Iguana iguana,is potentially due to a greater scope in both cardiac output and oxygen extraction (Gleeson et al.,1980; Bennett,1994). The iguana, at moderate to high speeds, is also unable to match the increased metabolic rate with adequate ventilation; as speed increases ventilation decreases (Carrier,1987a; Wang et al.,1997) and this decrease is accompanied with a decrease in V̇_O2(Wang et al., 1997). Such a finding adds support to the axial constraint hypothesis, at least for the iguana.

Despite the sustained interest in the determinants of aerobic scope in varanids, no study has yet measured all the O₂ transfer components in a single species. Further, the studies by Mitchell et al.(1981a,b)used arterial P_CO2(Pa_CO2) and pulmonary gas exchange rates to calculate ventilation and lung P_O2 from well-known alveolar gas equations. Such calculations assume that only ventilatory changes affect Pa_CO2.

This study therefore analyses O₂ transfer during sustainable treadmill exercise in a semi-aquatic varanid lizard, Varanus mertensi, to broaden our understanding of O₂ transfer beyond one varanid species. The use of a treadmill as opposed to a water channel is justified as this species is just as home on land as it is in water, and the use of a treadmill permits comparison with the results of previous studies.

Six water monitors Varanus mertensi Glauert 1951 (mass 1.39±0.39 kg, snout—vent length=40-45 cm) were obtained from the field (near Darwin, NT, Australia) and housed in temperature-controlled rooms(30 °C) with access to heat lamps. They were maintained on a diet of minced meat, insects and food supplement (Wombaroo, Australia), with water supplied ad libitum.

Animals were initially acclimated to the experimental temperature (35°C, body temperature, T_b, checked with an infrared thermometer) for at least 12 h before being run on a treadmill, set at speeds increasing from 0.2 to 0.42 m s^-1, and wearing a lightweight plastic mask that encompassed the head and enclosed the mouth and nostrils. A bias flow of air (2800 ml min^-1 STPD, standard temperature and pressure, dry) was drawn through the mask, dried and analysed for the rate of oxygen consumption, V̇_O2, as outlined below. Prodding of the tail ensured that the animals kept pace with the treadmill. The speed at which the maximum level of V̇_O2(V̇_O2max) was sustained for 2 min and showed no further increase with increasing speed was established as the speed to be used for subsequent experimental runs (0.33 m s^-1).

On a separate occasion, animals were fitted with Y-connector masks and ventilation and metabolism determined by the methods described below. The lizards were again run on a treadmill at 0.33 m s^-1 until exhaustion. These measurements were used to ensure that the subsequent surgery and placement of arterial and venous catheters did not affect the performance of the lizards.

Following the placement of catheters (see below), the animals were again fitted with Y-connector masks and made to run on the treadmill (0.33 m s^-1). This time ventilation and metabolism, end tidal gases (see below) and various blood parameters (see below) were measured at specified time intervals: pre-exercise (i.e. the animal masked and resting quietly on the treadmill for at least 30 min), during exercise at 1 min and 5 min,exhaustion (i.e. the animal no longer able to maintain pace), and after 2 min recovery.

The breathing pattern was analysed in terms of tidal volume(V_T), expiratory, inspiratory, inspiratory pause and total times (T_E, T_I, T_Pand T_TOT), frequency(f=1/T_TOT×60) and minute ventilation(V̇_E=V_T×f)in a similar approach to that detailed in Frappell et al.(1992). On average, 25 consecutive breaths were analysed. Volumes are expressed at BTPS(T_b, barometric pressure, saturated).

A 14-gauge needle was inserted from the centre of the two arms into the main stem of the Y-connector. This was connected directly to the other channel of the O₂ analyser and the CO₂ analyser with a short length of small-bore tubing (PE, i.d. 1.00 mm, o.d. 1.50 mm). Gas was subsampled through this circuit at a flow (τ=1.2 s) that permitted the determination of end tidal P_O2 and P_CO2 (PA_O2 and PA_CO2, respectively) [i.e. alveolar; strictly speaking the surface of the lung wall is increased by a honeycomb-like(faviform) system of partitions, which bear a matrix of capillaries on both surfaces. The individual chambers form the faveoli(Perry and Duncker, 1978)],determined as fractional concentration F×(PB-PH₂O₍₃₅₎), where PB is barometric pressure and PH₂O₍₃₅₎ is partial pressure of water vapour at 35°C. This gas then rejoined the main flow of air before it passed through the drying column.

For three animals the CO₂ analyser was placed in the subsampling circuit and used to determine end tidal P_CO2(PA_CO2); for the other three animals the CO₂ analyser was used to directly determine V̇_CO2.

Heart rate (fH) was determined from an electrocardiogram (e.c.g.)measured with two small needle electrodes placed on the dorsal service and positioned diagonally across the heart. The signal was appropriately amplified(7P4F, Polygraph, Grass Instruments).

Animals were anaesthetised with an i.v. dose of ketamine administered via the ventral coccygeal vein at a dose of 80-100 mg kg^-1. After induction of anaesthesia, a small mid-ventral incision(approximately 2 cm) was made in the neck. The common jugular vein was occlusively cannulated using a PE catheter (i.d. 0.58 mm, o.d. 0.96 mm). The tip of the catheter was directed past the brachiocephalic vein to rest in the anterior vena cava and was checked by measuring the length of the catheter on removal. The internal carotid artery was occlusively cannulated with PE catheter (i.d. 0.58 mm, o.d. 0.96 mm) and the tip advanced to rest as close to the heart as possible, presumably in the brachiocephalic artery or aorta. The catheters were sutured in place, looped loosely to avoid tension, led to the exterior through a small hole on the animal's dorsal surface just above the shoulders, filled with heparinized saline and sealed. The wound was closed with sutures. The surgical procedure lasted no more than 40 min and the animals were allowed to recover for at least 24 h prior to further testing.

Small blood samples (250-300 μl) were taken at designated times from both the artery (modifier, a) and the vein (mixed venous blood, modifier,v̄) and stored anaerobically on ice. The partial pressures of blood gases P_O2, P_CO2 and the pH were measured at 35°C (BMS MK2, Radiometer). The electrodes were calibrated before and after each measurement. P_O2 and P_CO2 were measured over 3 min and regressed back to time zero; pH was measured on incremental volumes of blood until the variation between successive measurements was less than 0.005 units. The oxygen content, C_O2, of each blood sample was determined on a 15μl subsample of blood using a galvanic cell (Oxygen Content Analyser, OxyCon). Hematocrit, hemoglobin concentration [Hb](Boehringer Mannheim kit no. 124729) and lactate concentration (Sigma kit no. 826) were also measured for each sample.

On another occasion a 2 μl blood sample was taken and oxygen equilibrium curves determined on a modified Hem-O-Scan (Aminco Instruments). The method followed is largely that presented by Holland et al.(1988), where gases of appropriate P_O2 are pulsed into the chamber and the percentage saturation determined. Curves were constructed from 10-15 points at 35°C and at CO₂ tensions of 39 and 79 mmHg (1 mmHg=133.3 Pa); the gases mixed by a pump (Wosthoff). Each point was digitised and the P₅₀ determined from the linear section of the Hill plot in the middle range of saturations. The Bohr factor was determined as(ΔlogP₅₀/ΔpH), pH being determined from blood tonometered at the P_CO2 values used in the Hem-O-Scan.

Resting lung volume and pulmonary diffusing capacity were determined for six animals after quietly resting for 30 min and during exercise using helium dilution and carbon monoxide clearance techniques. The lizards were fitted with a mouthpiece as previously described. The mask was directly attached via a three-way stopcock to a closed circuit of known volume(VSYS) that consisted of flexible tubing, helium and carbon monoxide analysers (CO and He Analysers, Morgan, UK), a collapsible reservoir and a pump that ensured a continual flow (approximately 1000 ml min^-1)through the circuit. The stopcock was initially open to enable the lizard to breathe room air. The circuit contained a test gas that comprised carbon monoxide (CO, 0.28%), helium (He, 0.138%) and air (the balance). During an end inspiratory pause, and with constant levels of He and CO in the circuit, the stopcock was turned so that the lungs were connected to the closed circuit. The breathing of the animal effected mixing of the test gas and alveolar gas. During this time the analysers continually monitored the levels of He and CO. The stopcock was opened to room air once the value of He reached a new steady state (Fig. 1).

A number of variables were derived.

All data signals were fed into an A/D board (DT2801/A, Data Translations,MA, USA) collecting at a speed of 50 Hz using ASYST (Macmillen Software,Keithley Instruments, NY, USA) and stored on computer for later analysis. Data are presented as means ± 1 S.D. Differences between treatments and time intervals were analysed using repeated-measures ANOVA. Post-hocmodified two-tailed t-tests were used to assess differences between appropriate comparisons using the Bonferroni method. A significant difference was defined as P<0.05.

No difference existed in V̇_O2 between the pre- and post-surgical treatments (P=0.254, Fig. 2). The pre-exercise V̇_O2(3.49±0.75 ml O₂ kg^-1 min^-1, Table 1) was higher than that reported for standard metabolic rate in this species(Christian and Conley, 1994),though this is hardly surprising given that the lizards in the present study were wearing a mask and rested for only 20 min during the day before running. V̇_O2 and V̇_CO2 increased with locomotion, quickly reaching a plateau that was maintained until exhaustion, on average 11.8 min after commencing the run(Fig. 2). Averaged over the period of exercise, V̇_O2max(Table 2) was similar to that previously reported (15.4±2.4 ml O₂ kg^-1min^-1; Christian and Conley,1994). R increased from 0.71±0.03 at rest to 0.86±0.21 during exercise. In the 2 min following exercise V̇_O2 tended to decline while V̇_CO2 was still maintained at the level achieved during running; as a result Rfurther increased (Table 2).

The breathing pattern at rest was typical for a lizard, i.e. evenly spaced single breaths, each interrupted by an end-inspiratory pause(Fig. 3). During sustained locomotion there was a substantial reduction in T_P (and a slight shortening of T_I and T_E, though this was not significant for T_I)(Fig. 3; Tables 1, 2). V̇_E increased(Table 2; Fig. 4), aided by increases in both V_T and f(Fig. 3B; Table 2), in direct proportion to V̇_O2(Fig. 4). Therefore, the air convection requirement for oxygen, V̇_E/V̇_O2,was maintained constant (Table 2; Fig. 4). A similar situation existed for V̇_E/V̇_CO2(Table 2). In the 2 min following exercise, V_T and f remained unchanged;as a result V̇_E was maintained at the exercise level and, as a consequence of the tendency for V̇_O2 to decline,the V̇_E/V̇_O2ratio tended to increase, though not significantly(Fig. 4). V̇_E/V̇_CO2was maintained at the same level as observed during exercise(Table 2).

The Hb affinity in V. mertensi(P₅₀=49.3±4.4 mmHg, 35 °C, P_CO2=39 mmHg, pH approx. 7.5), together with the Bohr factor (-0.30±0.048), were similar to values previously published in varanid species; the P₅₀ for V. exanthematicus is 47 mmHg at 35 °C, pH 7.5, Bohr factor -0.30(Wood et al., 1977), and for V. gouldii P₅₀ is 48 mmHg at 35 °C, pH 7.3(Bennett, 1973a). Despite a small increase in PA_O2 and Pa_O2 (Tables 1 and 2), the alveolar—arterial P_O2 difference(PA_O2—Pa_O2)remained unchanged during exercise from its value at rest (approx. 20 mmHg)and decreased slightly during the first few minutes of recovery(Fig. 4). Arterial O₂ content, Ca_O2, remained constant at all levels of V̇_O2, though due to the decrease in Pv̄_O2 the venous O₂ content, Cv̄_O2, dropped substantially (Tables 1 and 2). As expected, given no change in V̇_E/V̇_O2, Pa_CO2 remained constant and no different from PA_CO2 at all levels of V̇_O2 (Table 1 and 2). Despite a tendency for blood pH to decrease with exercise and during recovery this was not significant. Lactate concentration [La] increased approx. 2.5-fold during and after exercise (Tables 1 and 2). Hct and [Hb] remained constant throughout exercise and in the recovery phase; the value of Hct was identical to that reported for V. gouldii(Bennett, 1973a).

The associated increase in arterial—venous O₂ content difference(Ca_O2—Cv̄_O2)with exercise was not enough to offset any increases in cardiac output, hence Q̇_tot increased (Tables 1 and 2). The increase in Q̇_tot was achieved through a corresponding increase in fH, and stroke volume remained unaltered(approximately 1.72 ml kg^-1). Both Q̇_tot and fH remained high during the first 2 min of recovery. At rest V. mertensipossessed a considerable right-to-left shunt(Table 1), but unfortunately no information was obtained on the shunt during exercise.

Locomotion had no effect on lung volume (VL) or the pulmonary diffusing capacity for carbon monoxide (DL_CO) (Table 1 and 2).

Not all varanids posses a high aerobic scope, and among their reptilian counterparts an increasing body of evidence is beginning to show that aerobic scope in this group is as varied as it is amongst mammals(Christian and Conley, 1994). The values reported for V̇_O2max in this study for lizards either wearing the mask that encompassed the head or fitted with the mouthpiece (and pre- or post-surgery) were the same, and similar to that previously reported for this species(Christian and Conley, 1994). This value is approximately 20 % lower than that predicted for a similarly sized varanid by the interspecific equation for varanid V̇_O2max(Thompson and Withers, 1997),and is possibly related to the semiaquatic lifestyle of this species.

The [La] values reported here (8.5 mmol l^-1) are similar to levels found in V. exanthematicus (7.2 mmol l^-1) at maximal sustainable speed (Gleeson et al.,1980). Both these values are much lower than the values(approximately 28 mmol l^-1) associated with high and probably unsustainable levels of activity in that species(Mitchell et al., 1981a). Obviously, accumulation of [La] implies that anaerobic metabolism is supplementing aerobic energy production. During intensive bouts of activity,anaerobic glycolysis in reptiles can account for most of the ATP produced,with aerobic metabolism playing a minor role(Bennett, 1994). The reader is referred to the reviews by Gleeson(1991, 1996); also see(Bennett, 1994) for a detailed account of anaerobic metabolism during intense activity in reptiles and the aerobic payback that accompanies recovery.

Of particular interest in this study, given the recent findings that gular movements may augment ventilation in V. exanthematicus(Owerkowicz et al., 1999), are the similar values in V̇_O2 achieved for lizards exercised with a mask encompassing the head or while fitted with a mouthpiece. In the study by Owerkowicz et al.(1999), lizards in which the mouth was propped open were unable to reach the level of V̇_O2max achieved when exercised with their mouth closed and therefore capable of gular pumping. The implication from that study was that gular pumping was required to achieve appropriate levels of V̇_Enecessary to support the high levels of V̇_O2 associated with exercise. If the presence of a functional gular pump is an attribute of all varanids during exercise, then it may be concluded that the mouthpiece used in the present experiments did not impair the gular pump. However, we did not measure buccal cavity pressure to determine if gular pumping was occurring. On the other hand, Schultz et al.(1999) studied four species of varanids (V. spenceri, V. gouldii, V. panoptes and V. mertensi) and showed that, whereas at rest all species relied on nasal breathing, when exercised they did so with their mouths open to various degrees, which could account for 11-49 % of the expired gases. This finding tends to suggest that gular-assisted ventilation is not obligatory and/or widespread amongst the varanids. Obviously, further studies are required to assess the degree to which gular pumping is a characteristic of varanids.

The finding that varanids breathe through their mouths during exercise raises some concerns about the levels of ventilation reported in previous studies, where ventilation was measured only through the nostrils. Carrier(1987a) assessed inspiratory airflows through the nostrils with a thermistor, and Hopkins et al.(1995) and Wang et al.(1997) measured airflow through a pneumotacograph attached to the nostrils. Hopkins and coworkers checked for mouth breathing at rest in an attempt to validate this approach and found none; perhaps not surprising, given the findings of Schultz et al.(1999).

It has been hypothesised that mechanical interference exists between locomotor and ventilatory function in lizards, as the intercostal muscles are required for both locomotion (lateral flexion of the trunk) and ventilation(expansion of the thorax) (Carrier, 1987a, b, 1989, 1990). Accordingly, the aerobic needs of an exercising lizard cannot be met because of the locomotory constraints on lung ventilation. Indeed, Wang et al.(1997) found no change in tidal volume during exercise in V. exanthematicus but a two- to fourfold increase immediately following exercise. An increase in V̇_E did occur, however,during exercise as a result of changes in f. The present study with V. mertensi also finds an increase in V̇_E during exercise, the result of an increase in V_T and f (primarily through a shortening in T_P), and this increase is retained in the recovery period immediately following exercise. As previously mentioned, costal ventilation during exercise was most likely constrained in V. exanthematicus and gular pumping, by circumventing the constraint,contributed to the overall increase in ventilation and hence V̇_O2(Owerkowicz et al., 1999). Gular pumping is not used in Iguana iguana and ventilation decreases at speeds greater than a very slow walk(Wang et al., 1997; Owerkowicz et al., 1999).

While costal ventilation may be inhibited in some lizards during exercise,the undulatory movement of the body trunk that accompanies locomotion may be of benefit to gas exchange. The highly compliant nature of the respiratory system in lizards (Perry and Duncker,1978; Milsom,1989) provides minimal constraint to changes in body posture and is prone to distortion. Any distortion of the lungs that occurs could enhance the effectiveness of gas exchange by improving gas mixing. However, in an experiment with V. gouldii, where individuals were hyperventilated to lower the Pa_CO2 levels below the apneic threshold, no change was detected in the time course of arterial blood gas levels with or without lateral passive body movements(Frappell and Mortola, 1998). This suggested that lateral chest wall movements neither hindered nor facilitated gas exchange.

In the study of Wang et al.(1997), the air convection requirement(V̇_E/V̇_O2)in V. exanthematicus approximately doubled during exercise at aerobically sustainable speeds, and on recovery returned to pre-exercise levels, indicating a substantial hyperventilation with exercise. In contrast,the present study finds that the increase in V̇_E during exercise in V. mertensi adequately meets the increase in V̇_O2, hence maintaining the V̇_E/V̇_O2ratio constant between exercise and pre-exercise. While the study of Bennett(1973b) reports a relative hypoventilation for V. gouldii immediately after activity, the results should be viewed with some degree of scepticism, as activity was achieved following electrical stimulation of restrained lizards.

While not directly measuring ventilation, Mitchell and coworkers(Mitchell et al., 1981b)indicated, from measurements of Pa_CO2 and the alveolar gas equation, that during exercise V. exanthematicusincreased V̇A relatively more than V̇_O2, giving rise to an increase in PA_O2 (again calculated, not measured). Likewise, Hopkins et al.(1995) report a small hyperventilation in V. exanthematicus, though the associated increase in Pa_O2 was not significant. The present study also revealed a small increase in PA_O2 and Pa_O2 with exercise and recovery in V. mertensi. In the absence of any change in V̇_E/V̇_O2,the increase in PA_O2 suggests a proportionally greater increase in alveolar ventilation (i.e. relative decrease in dead space). At rest, V̇A accounts for approx. 70 % of V̇_E, whereas during exercise and recovery this increased to approx. 95 % (Tables 1 and 2). In summary, the available evidence suggests that, for varanids exercised at sustainable speeds, V̇_E increases sufficiently during exercise to meet the increased V̇_O2, and the small increase in PA_O2 observed most likely reflects a proportionally greater increase in V̇A.

At rest, a substantial alveolar—arterial P_O2(PA_O2-Pa_O2)difference was observed in V. mertensi (approx. 21 mmHg), as has been previously reported for V. exanthematicus (approx. 14 mmHg, Mitchell et al., 1981b;approx. 27 mmHg, Hopkins et al.,1995). During exercise the increase in PA_O2 and Pa_O2 in both species (this study; Hopkins et al.,1995) maintains PA_O2-Pa_O2,whereas Mitchell et al.(1981b) reported an increase in PA_O2-Pa_O2during exercise to values similar to that found by Hopkins and coworkers. The level of PA_O2-Pa_O2 is an indication of the efficacy of alveolar—arterial O₂ transfer and could result from right-to-left (R—L) shunts, diffusion limitation or ventilation—perfusion inhomogeneity.

At rest reptiles may possess R—L shunts. A substantial R—L shunt of between 16 % (Berger and Heisler,1977) and 6 % (Hopkins et al.,1995) has been reported in V. exanthematicus at rest,though the lower value reported by Hopkins is for an intrapulmonary shunt only, as their measurements excluded intracardiac shunting. The present study reports a R—L shunt in V. merstensi of approx. 17 % at rest. Studies on pulmonary gas transfer must, strictly speaking, be based on blood gases sampled from the pulmonary veins(Wang et al., 1998). In the present study our measure of Pa_O2 could potentially include intracardiac shunts, which conceptually are identical to that of intrapulmonary shunts. However, based on studies on V. niloticus (Millard and Johansen,1974; Ishimatsu et al.,1988) and V. exanthematicus(Burggren and Johansen, 1982),it would appear that in varanids R—L intracardiac shunting is much reduced or absent.

During exercise V. exanthematicus decreased the R—L intrapulmonary shunt from about 6 % to 2 %(Hopkins et al., 1995). Further, many reptiles develop a significant L—R shunt during exercise(Hicks and Krosniunas, 1996),though in V. exanthematicus this was shown not to occur(Hopkins et al., 1995). A reduction in the R—L shunt may improve Pa_O2, depending on V̇_O2 and blood O₂ carrying capacity. If the O₂ levels of the pulmonary veins are on the flat portion of the oxygen dissociation curve then Ca_O2 and Pa_O2will be affected only marginally (Wang and Hicks, 1996). Further, changes in Ca_O2-Cv̄_O2can compensate for changes in shunt fraction to produce the same Pa_O2 (see Equation 9 used for determining Q̇_shunt/Q̇_tot).

Compared with those of mammals, reptilian lungs have a lower surface area for diffusion and a blood gas barrier of increased and variable thickness(Perry, 1983; Perry et al., 1994). From the relationship between V̇_O2 and DL_O2[V̇_O2=DL_O2×(PA_O2-Pa_O2)],the alveolar—arterial difference (for reptiles,falveolar—pulmonary vein difference) can be determined. While reptiles have a lower metabolic rate than mammals, DL_O2is even lower, and as a result the equivalent PA_O2-Pa_O2 will be higher in reptiles than mammals (Glass,1991). The present study measured DL_CO(DL_O=1.23DL_CO) in V. mertensiand found at rest a value about double that previously reported in V. exanthematicus at 35°C (Glass et al., 1981). In V. exanthematicus it was concluded that diffusion limitation did not contribute significantly to the PA_O2-Pa_O2 at rest (Hopkins et al., 1995)and presumably, given a higher DL_CO, the same also applies to V. mertensi. During exercise, however, Hopkins and coworkers found that diffusion limitation accounts for a substantial part of the PA_O2-Pa_O2difference in V. exanthematicus, a situation often reported in mammals (see Powell,1994).

The overall ventilation perfusion ratio for the lung was 0.71 at rest and increasing to 2.13 during exercise in V. exanthematicus, similar to that estimated{V̇E/Q̇L=V̇A/[Q̇_tot×(1-Q̇_shunt/Q̇_tot)],where Q̇L is the blood flow through the lung} for V. mertensi (approx. 0.8 and 2.1, respectively). With exercise this reflects, in both species, a greater increase in ventilation than perfusion. Alveolar—arterial P_O2difference is also enhanced by spatial ventilation—perfusion inhomogeneity, which during exercise increased in V. exanthematicus(Hopkins et al., 1995). A similar situation occurs in exercising mammals, though the mechanisms responsible for the development of ventilation—perfusion inhomogeneity during exercise remain unclear (Wagner et al., 1986; Schaffartzik et al., 1992; Hopkins et al.,1994). Together, the available information suggests that pulmonary gas exchange in reptiles is no more limiting in varanids than it is in mammals.

Oxygen delivery to the tissues depends on both Q̇_tot and Ca_O2-Cv̄_O2(oxygen extraction). The reliance on both Q̇_tot and oxygen extraction by the tissues during exercise is fairly common amongst vertebrates(Gleeson et al., 1980) and occurred in V. mertensi; the increase in V̇_O2 during exercise is achieved through increases in both O₂ extraction and cardiac output (Fig. 5). As PA_O2 and Pa_O2are maintained or slightly increased during exercise, high values of arterial saturation were preserved (Fig. 5). Likewise, V. exanthematicus had a high Ca_O2 which, in turn, permitted a greater Ca_O2-Cv̄_O2during sustainable levels of exercise(Gleeson et al., 1980). Wood et al. (1977) also note that the arterial—venous O₂ content difference is large in V. exanthematicus compared with most other groups of reptiles, and that this was a result of low shunt flows and corresponding high values of arterial saturation (Pa_O2=94 mmHg, at 35°C), hence high Ca_O2.

The ability to maximise oxygen extraction during exercise is assisted by a distinct Bohr shift in the hemoglobin saturation curve that further favours unloading of O₂, and a greater increase in Ca_O2-Cv̄_O2than would otherwise be achieved (Fig. 5). Previously, Gleeson et al.(1980) noted that both V. exanthematicus and I. iguana decreased mixed venous blood O₂ content to similar levels; the difference in Ca_O2-Cv̄_O2was a reflection of the varanid's greater Ca_O2. Further, in that study, while the factorial scope in cardiac output was greater in V. exanthematicus, both the varanid and iguana reached maximum Q̇_tot at intermediate levels of V̇_O2. This implies that without further increases in O₂ extraction,transport by the circulation could be limiting. It is also interesting to note that although the air convection requirement(V̇E/V̇_O2)remained constant, the blood convection requirement(Q̇_tot/V̇_O2)decreased with exercise in V. mertensi, V. exanthematicus(Gleeson et al., 1980; Wang et al., 1997) and I. iguana (Gleeson et al.,1980). A decreasing blood convection requirement during exercise and an already large arterial—venous O₂ content difference suggests that there exists limited capacity for further increases in O₂ transport by the circulatory system. More recently, Farmer and Hicks (2000) suggested that in I. iguana both the circulatory and ventilatory systems imposed limits to O₂ transfer at anything faster than a slow walk. However, in the iguana maximum blood flow and ventilation occur during recovery(Wang et al., 1997; Farmer and Hicks, 2000). This is in contrast to varanids, where ventilation and cardiac output reach a maximum during exercise (Gleeson et al.,1980; Wang et al.,1997; this study).

In conclusion, we have examined O₂ transfer in the varanid, V. mertensi, during maximal levels of sustained aerobic exercise. Ventilation and O₂ transfer across the lung would appear to be adequate to meet the aerobic needs of V. mertensi during exercise. This is also the case in mammals, where the lungs are built with a significant excess structural capacity (Hoppeler and Weibel, 1998). On the other hand, at V̇_O2max the circulatory system in mammals operates at or close to the upper limit of structural capacity for O₂ transport(Hoppeler and Weibel, 1998). Together with an already large arterial—venous O₂ content difference in V. mertensi, a decreasing blood convection requirement with exercise suggests a physiological limit occurring with the transport of oxygen by the circulatory system.

List of symbols used
 
• ā

  arterial modifier

 
• C_O2

  blood oxygen content

 
• Ca_O2

  arterial oxygen content

 
• Cc′_O2

  O₂ content of end-capillary blood

 
• C v̄ _O2

  mixed venous O₂ content

 
• DL

  diffusing capacity of the lungs

 
• E

  excurrent modifier

 
• f

  breathing frequency

 
• fH

  heart rate

 
• F_CO

  fractional concentration of CO

 
• F_CO2

  fractional concentration of CO₂

 
• F_O2

  fractional concentration of O₂

 
• Hb

  haemoglobin

 
• Hct

  haematocrit

 
• I

  incurrent modifier

 
• La

  lactate

 
• P₅₀

  P_O2 at which haemoglobin is 50 %saturated

 
• Pa_CO2

  arterial partial pressure of carbon dioxide

 
• Pa_O2

  arterial partial pressure of oxygen

 
• PA_O2

  alveolar partial pressure of oxygen

 
• PB

  barometric pressure

 
• P_CO2

  partial pressure of carbon dioxide

 
• PH₂O(35)

  partial pressure of water vapour at 35°C

 
• P_O2

  partial pressure of oxygen

 
• P v̄ _CO2

  partial pressure of mixed venous carbon dioxide

 
• P v̄ _O2

  partial pressure of of mixed venous oxygen

 
• Q̇ _L

  blood flow through the lung

 
• Q̇ _shunt

  shunt blood flow

 
• Q̇ _tot

  total cardiac output

 
• R

  respiratory exchange ratio

 
• R

  gas constant

 
• T

  absolute temperature

 
• t

  time

 
• T_b

  body temperature

 
• T_E

  expiratory time

 
• T_I

  inspiratory time (does not include T_P)

 
• T_P

  inspiratory pause

 
• T_TOT

  total breathing time

 
• v̄

  mixed venous blood, modifier,

 
• V̇ _A

  alveolar ventilation

 
• V_AIR

  volume of air

 
• V̇ _CO2

  rate of carbon dioxide production

 
• V̇ _E

  minute ventilation

 
• V_L

  volume of the lung

 
• V̇ _O2

  rate of oxygen consumption

 
• V̇ _O2max

  maximum rate of oxygen consumption

 
• V_R

  ventilation at rest

 
• V_SYS

  closed circuit of known volume

 
• V_T

  tidal volume of breathing

 
• βO₂

  capacitance of blood for oxygen

 
• τ

  gas flow

This work was conducted under relevant Animal Ethics Committee and wildlife authority permits. Gavin Bedford, Gary Smith, Kristin Munro and James Boags are gratefully acknowledged for assistance in the field and/or with animal care. This project was supported by funds from the Australian Research Council to P.B.F. and K.A.C. and contributions from La Trobe and Northern Territory Universities. T.J.S. was in receipt of an Australian Postgraduate Research Award.