The Role of the Cardiac Vagus in the Response of the Dogfish Scyliorhinus Canicula to Hypoxia

For effective respiratory gas exchange it is important not only to have a suitable relationship between blood flow and the flow of the respiratory fluid, but also that this relationship should be capable of controlled alteration in response to changes in oxygen demand or the relative tensions of the respiratory gases (Jones, Randall & Jarman, 1970). The reflex response of a number of teleosts to a reduction in environmental oxygen tension, for example, is a rise in ventilation volume and a reduction in heart rate (Randall & Shelton 1962; Holeton & Randall, 1967 a, b; Marvin & Heath, 1968). Elasmobranchs are also known to exhibit a bradycardia in response to hypoxia (Satchell, 1961; Piiper, Baumgarten & Meyer, 1970; Butler & Taylor, 1971, 1975). These changes in heart rate have been shown to be mediated via the cardiac branches of the vagus in teleosts (Randall 1966; Randall & Smith, 1967).

The elasmobranchs are unusual amongst vertebrates in that efferent nervous control of heart rate appears to be entirely dependent on variations in the level of tonic inhibition mediated by the parasympathetic vagus (Lutz, 1930; Young 1931, 1933; Gannon, Cambell & Satchell, 1972; Short, 1976). Variations in heart rate in response to hypoxia are, therefore, completely abolished by the administration of atropine which effectively blocks cholinergic synapses (Butler & Taylor, 1971).

Innervation of the heart in elasmobranchs is also unusual in that it consists of two distinct pairs of cardiac branches from the vagus; the visceral cardiac branches which arise from the visceral branches of the vagus (Marshall & Hurst, 1905) and the branchial cardiac branches which arise from the 14th branchial branches of the vagus (Norris & Hughes, 1920).

The present investigation sets out to determine the role of the cardiac vagus in controlling heart rate, and particularly the bradycardia during hypoxia, the intensity of which varies with temperature (Butler & Taylor, 1975). The influence of the vagus on heart rate and blood flow in the ventral aorta was investigated at a range of temperatures both pharmacologically and by selective sectioning of the two pairs of cardiac nerves, together and separately, in order to determine their combined and separate roles.

The present report describes results from five series of experiments on a total of 79 dogfish, Scyliorhinus canicula L., of either sex whose mass varied between 0·3 and 1·2 kg. In the first series, 14 dogfish of between 0.5 and 1 kg were obtained from the laboratories of the Marine Biological Association, Plymouth, and maintained at acclimation temperatures of either 7 °C (six animals) or 17 °C (eight animals) for a minimum of 2 weeks prior to experiments. The experimental regime was identical to that used previously on dogfish acclimated to 12 °C (Butler & Taylor, 1971) and the present results are compared with this earlier study. Briefly the fish were anaesthetised in MS 222 (Sandoz) whilst cannulae were placed in the orobranchial cavity and dorsal and ventral aortae. Heart rate, respiratory frequency, blood pressure, inspired and arterial were measured from animals in normoxic sea water and during gradually induced hypoxia over a period of 1 h both before and after injection of 0·15 mg kg⁻¹ of atropine sulphate (Sigma Chemical Co.) via a cannula in the caudal vein. This initial series of experiments confirmed the observations by Butler & Taylor (1975) that the responses of the dogfish to hypoxia were affected by temperature. For the remainder of this investigation the acclimation temperature of the fish was selected in order to obtain either a relatively pronounced or a slight cardiac chronotropic response to hypoxia, as required.

The second experimental series was performed on 30 dogfish of either sex whose mass varied between 0·30 and 1·00 kg, which were trawled from Carmarthen Bay acclimated for 2 weeks to 17 °C in large holding tanks of recirculated and aerated sea water, and experimented upon in the Department of Zoology at the University of Swansea. The experimental technique was similar to that described by Butler & Taylor (1971) and the variables measured were respiratory frequency, heart rate ⁠, and the blood pressures in the dorsal and ventral aortae. In this experimental series the influence of the cardiac vagus on heart rate was removed by selective nerve transection, prior to atropinisation. The innervation of the heart and branchial arches was traced by dissection of six freshly killed dogfish and a diagram of the anatomy of this region is shown in Fig. 1. The identification of the cardiac nerves was checked by electrical stimulation. This study of the anatomy of the relevant parts of the nervous system enabled the cardiac nerves to be sectioned where they were most easily accessible and was also used to identify the cardiac nerves for electrical stimulation (see Short, Butler and Taylor, 1977) and the branchial nerves for selective transection (Butler, Taylor & Short, 1977).

The branchial cardiac nerves were sectioned by cutting the branchial branches of the vagus to the last (5th) gill arch on either side of the fish. The visceral cardiac nerves were sectioned at their point of origin on the visceral branch of the vagus (see Fig. 1). This procedure allowed both pairs of nerves to be sectioned, together or separately, without entering either anterior cardinal sinus. Blood loss during the operation was, therefore, reduced to a minimum and the possibility of air entering the blood system was avoided. Following nerve section the wounds were sutured and the fish were replaced in holding tanks for 2–3 days before being re-anaesthetized and prepared for experiments.

The fish used in this series of experiments were divided randomly into the following five control and experimental categories:

• control fish ; those which did not have their vagi exposed ;

• sham-operated fish; those in which the cardiac vagi were exposed as if for sectioning but left intact;

• branchial cardiac vagotomized fish; those with the branchial cardiac nerves sectioned bilaterally;

• visceral cardiac vagotomized fish ; those with the visceral cardiac nerves sectioned bilaterally;

• totally cardiac vagotomized fish ; those with both pairs of cardiac vagi bilaterally sectioned.

In this second series of experiments the fish were exposed to rapidly induced hypoxia (see Butler & Taylor, 1971). The was reduced from the normoxic value of approximately 135 mmHg down to less than 30 mmHg within 1 min.

In the remaining three series of experiments the changes in blood flow in the ventral aorta in response to hypoxia were measured directly using an electro-magnetic flow probe (Biotronex Ltd). A cannulating electro-magnetic flow probe was inserted into the portion of ventral aorta supplying the first two pairs of afferent branchial arteries. At the end of the experiment the probe was removed from the fish and calibrated with the fish’s own blood at the experimental temperature. The measured flows obtained by this technique are referred to as stroke flow or total flow rather than stroke volume or cardiac output as it is not possible to measure the total output of the heart using this technique. A hole in the lumen of the probe was connected to a cannula and in addition a cannula was inserted into the dorsal aorta. These cannulae were used to measure blood pressure and enabled blood to be sampled. The techniques involved are described by Butler & Taylor (1975). These experiments were performed on a total of 29 fish of either sex whose mass varied between 0·7 and 1·2 kg, which were obtained from Plymouth and acclimated to 15 °C. Use of these direct measurements of blood flow to describe the changes consequent on exposure to hypoxia is only valid if the proportion of total cardiac output flowing via the flow probe to the first two gill arches does not change. To test this the blood flow to the first two gill arches was measured directly on four dogfish which were acclimated to 15 °C and confined in a continuous flow respirometer (see Butler & Taylor, 1975). Samples of arterial and venous blood were also taken to enable blood flow (i.e. cardiac output) to be estimated using the Fick principle. Blood oxygen tensions and were measured using a blood gas analyser (Radiometer PHM 71), and the electrode was calibrated with saline equilibrated with N₂or air at the experimental temperature. Blood oxygen contents and were measured directly using a Lex O₂Con (Lexington Instruments), and haematocrit was measured using a microhaematocrit centrifuge (Hawkesley). Of the remaining 25 fish acclimated to 15 °C, ten were exposed to both rapidly induced and gradual hypoxia (see Butler & Taylor, 1971) whilst in a ‘control’ condition (i.e. prior to atropinization or exposure of the cardiac vagus). Following this initial exposure to hypoxia these fish were atropinized then exposed once more to hypoxia. The means of the measured variables from these two consecutive experiments on individual fish were compared statistically.

In seven fish the four cardiac nerves were dissected away from surrounding connective tissue and a 20 cm length of silk suture thread was fed beneath the nerve trunk; both ends of this thread were then pushed through a 5 cm length of polythene cannula to form a snare. The four snares were sewn into the wound so that they protruded through the skin covering the anterior cardinal sinus and the fish were allowed to recover from the anaesthetic. As the cardiac nerves of these fish had been exposed but not sectioned they provided a control for the two techniques and are referred to as ‘sham-operated’ fish. Following exposure to hypoxia the four nerves were sectioned by pulling the snares and after 1 h recovery the fish were exposed once more to hypoxia. The measured variables following vagotomy were compared statistically with the values obtained from the sham-operated fish. These experiments allowed the effects of atropinization on heart rate and blood flow to be compared with the effects of total cardiac vagotomy.

In the remaining eight fish acclimated to 15 °C, blood flow in the ventral aorta and pressures in the dorsal and ventral aortae and central venous system were measured (see Short et al. 1977), and the cardiovascular responses to hypoxia were observed before and after injection of the adrenergic β-receptor blocking agent Propranolol, through the cannula in the caudal vein. It was found that 0·4 mg kg⁻¹ of DL-propranolol HCI (Sigma) was just sufficient to abolish the cardiovascular effects of an injection of 0·1 mg kg⁻¹ of the pure adrenergic beta-receptor stimulating agent DL-isoproterenol sulphate (Sigma), which was found to be the threshold concentration for a clear effect on the cardiovascular system of control fish. This concentration of the antagonist was, therefore, adopted routinely and its efficacy was checked by injection of the agonist at the conclusion of each experiment.

Finally, blood flow in the ventral aorta of six control fish (mean mass 0·81 ± 0·04 kg) was measured during exposure to gradual hypoxia down to a of 25 mmHg in five steps each of 30 min duration, as described by Butler & Taylor (1975).

In all experiments the fish were disturbed as little as possible and were allowed at least 3 h to recover from the effects of the MS 222 before any physiological variables were measured. MS 222 anaesthesia is known to cause a rise in blood ⁠, lactate fed haematocrit and a reduction in and and pH_a in the rainbow trout, but all these variables recover back to their normal, unanaesthetized levels within 1 h (Soivio, 1976),which is the time taken for the anaesthetic to clear from the blood (Houston & Woods 1972). Diluted blood samples were observed in a scanning spectrophotometer and it was found that prolonged (2 h) exposure to MS 222 caused a reduction in the oxyhaemoglobin peak. This effect could have resulted from the production of methaemoglobin as the effect was mimicked by diluting the blood with 0·01 M potassium ferricyanide solution. The peaks had, however, returned to normal within 1 h.

Unless otherwise stated, all numerical values in the test are given as mean values ± S.E. of mean. Student’s t test was used to test the significance of any difference between two mean values and the word ‘significant’ in the present report means that the difference is significant at the 95% confidence level (P<0·05). The description ‘control fish’ in the present report refers to post-operative dogfish under the experimental conditions described, in which the nervous control of the heart remained intact (i.e. prior to atropinization or vagotomy).

When dogfish were exposed to a gradual decrease in from above 135 mmHg to below 35 mmHg at 7, 12 or 17 °C, their responses were identical to those previously described by Butler & Taylor (1971, 1975). The effects of temperature on respiratory frequency, heart rate and blood pressure have been discussed previously (Butler & Taylor, 1975). Blood pressure and respiratory frequency were unaffected by injection of atropine. The present investigation concentrates, therefore, upon the cardiac chronotropic responses to atropinization and to hypoxia, plus the associated changes in ⁠. These variables are listed in Table 1.

During normoxia heart rate varied in response to both acclimation temperature and injection of atropine. At 7 °C injection of atropine had no significant effect on heart rate (P>0·1) whereas at 12 and 17 °C there were significant rises in cardiac frequency following atropinization (P< 0·02 and <0·01 respectively). At 12 °C normoxic heart rate rose to 118 ± 8% of the control value and at 17°C rose to 127 ± 6% of the control value following the injection of atropine.

During gradual hypoxia there was a slight reduction in heart rate in the control dogfish acclimated to 7 °C and at a of 30 mmHg heart rate was 90 ± 4·5 of the initial rate, this reduction was not, however, significant. At 12 and 17 °C there were significant reductions in heart rate during hypoxia. At a of 30 mmHg heart rate was 76 + 5% of the initial rate at 12 °C and 46 ±2·5% of the initial rate at 17 °C. The injection of atropine abolished the bradycardia in response to gradual hypoxia at all three temperatures (see Table 1).

There was, of course, a progressive reduction in during gradual hypoxia. Atropinization had no significant effect on the hypoxic value at 7 °C. At 12 and 17 °C there was a progressive trend for hypoxic to be lower following atropinization, when compared at similar levels of ⁠. The mean values for hypoxic before and after the injection of atropine, shown in Table 1, were significantly different at 17 °C (P<0·05).

Mean values of respiratory frequency, heart rate, dorsal and ventral aortic blood pressure, and were measured on control fish, on sham-operated fish and on fish following selective transection of the branchial cardiac and visceral cardiac branches of the vagus as described above. The variables were measured first in normoxic sea water (the initial values) then during exposure to rapid hypoxia when the was reduced to less than 35 mmHg within 1 min (see Fig. 2), and finally on recovery in normoxia. Following a 3 h period for recovery each fish was atropinized and exposed once again to hypoxia. Prior to atropinization, and with the fish in normoxic sea water, heart rate was the only variable to show significant differences between treatments. The heart rates of the control (38·5 ± 2·6 (6) beats min⁻¹), sham-operated (39 ± 3’2 (6) beats min⁻¹) and visceral cardiac vagotomized (39·3 ± 3·0 (7) beats min⁻¹) fish were statistically similar. Both the branchial vagotomized and the totally cardiac vagotomized fish had mean normoxic heart rates which were significantly higher than in the control fish. The heart rate of the branchial vagotomized fish (44·9 ± 1·0 (7) beats min⁻¹) was lower than that of the totally cardiac vagotomized fish (47·3 ± 1·3 (4) beats min⁻¹), but this difference was not significant (P > 0·05). Following atropinization each of the measured variables in normoxia, including heart rate, from all the five categories were statistically similar. Comparison of the mean values for variables within each category revealed that normoxic heart rate increased significantly following the injection of atropine in all but the totally vagotomized fish. The percentage increases were 28% in control animals, 18% in sham-operated, 23% in visceral cardiac vagotomized animals, but only 10% in branchial cardiac vagotomized animals. These increases are apparent in Fig. 3.

The cardiac chronotropic responses of the fish to rapidly induced hypoxia are shown in Figs. 2 and 3. During the rapid reduction in the heart rate of control fish fell from the initial normoxic rate down to a transient low rate often within one beat (see Fig. 2). This transient rate was computed from the longest cardiac interval measured during the instantaneous response to rapid hypoxia. Following this instantaneous intense bradycardia, heart rate recovered to a stable rate characteristic of the measured which was approximately 49 % of the initial, normoxic rate. An arterial blood sample was withdrawn during this period for measurement of the hypoxic When the water was reaerated, heart rate recovered rapidly towards its initial rate (see Fig. 2). Heart rate during recovery from hypoxia was often higher than the initial rate (see Fig. 3), indicating a transient recovery tachycardia. The increase in mean values was, however, not significant. Following atropinization there was no change in heart rate during rapid hypoxia. This pattern of changes in the heart rate of control fish at the onset of and during recovery from rapidly induced hypoxia was described and discussed in some detail by Butler & Taylor (1971).

The changes in heart rate of the control, sham-operated and visceral cardiac vagotomized fish during rapid hypoxia followed the same pattern, i.e. the initial, transient, stable, recovery and atropinized heart rates were each statistically identical between these three treatments. Within each of these experimental categories the initial, transient, stable and atropinized heart rates were all significantly different from one another (see Fig. 3). The totally cardiac vagotomized fish showed no changes in heart rate during rapid hypoxia. The branchial cardiac vagotomised fish showed a modified response to rapid hypoxia; instead of a transient low heart rate at the onset of hypoxia, these fish showed a slowly developing bradycardia down to a stable rate which was significantly higher than in the control animals, being approximately 73% of the initial rate. This modified response to rapid hypoxia is illustrated in Figs. 2 and 3.

As found in the first series of experiments there was always a significant reduction of during hypoxia, both before and after atropinization. The mean during hypoxia was significantly reduced from 15·2+1·5 (6) down to 9·7 ±0·7 (6) mmHg in control animals following atropinization (P<0·05) (Table 1) and similar reductions were also apparent in sham-operated and visceral cardiac vagotomized fish, though they failed to satisfy the t test criteria (P > 0·05). The hypoxic values measured from branchial cardiac vagotomized and totally cardiac vagotomized fish prior to atropinization were significantly lower than the control and sham-operated values but statistically similar to the values measured following the injection of atropine (see Fig. 3).

Blood flow through a section of the ventral aorta as well as blood pressure, arterial and venous blood oxygen tensions and contents and were measured simultaneously in four fish acclimated to 15 °C both in normoxia and following exposure to gradually induced hypoxia down to a of 60 mmHg, where they were held for 1 h prior to the samples being taken. Total cardiac output (Q) was derived from these measurements via an equation based on the Fick principle (see Butler & Taylor, 1975), and compared with the direct measurements of stroke flow in the anterior portion of the ventral aorta. The results are presented in Table 2.

During hypoxia the heart rate fell to 46% of its normoxic level and the stroke volume increased to 136% of the normoxic level. This resulted in a decrease in the derived value for cardiac output to 61 % of the normoxic level. The direct measurements of stroke flow changed in a similar way during hypoxia, and the proportion of total cardiac output directed to the anterior two pairs of afferent branchial arteries, via the implanted flow probe, remaining constant at approximately 37%. The important feature of this result was that it established the validity of the direct measurements of stroke flow used for the remainder of this investigation.

The mean normoxic values of the measured variables in control fish were compared with those obtained following atropinization, exposure of the four branches of the cardiac vagus (i.e. the ‘sham’ operation) and finally total cardiac vagotomy. There were no significant differences between the mean normoxic values of the measured variables from normal and sham-operated animals with the exception of mean dorsal blood pressure, which was significantly lower in the sham-operated fish. This may be a result of the exposure of the cardiac nerves prior to sectioning. The normoxic heart rate increased to 122% of the level in control animals following atropinization (from 27·3±1·8(10) to 33·2±1·0(10) beats min⁻¹) and to 130% of the level in sham-operated fish following cardiac vagotomy (from 26·1 ±3·4 (7) to 34·4 ± 0·8 (7) beats min⁻¹). Each of these increases in heart rate was associated with a reduction in stroke flow, though this was only significant in the cardiac vagotomized fish (from 0·32 ±0·05 to 0·20 ±0·03 (7) ml kg⁻¹). The normoxic values for all other measured variables did not differ significantly between the 4 treatments.

Following exposure to hypoxia down to a of approximately 30 mmHg, heart rate was reduced to 60 % of the normoxic rate in control animals and to 69% of normoxic rate in sham-operated animals. These reductions in rate were accompanied by respective increases in stroke flow to 127% and 123 % of the normoxic values. The increase in stroke flow compensated for the bradycardia to the extent that there was no significant reduction in total blood flow during hypoxia in sham-operated fish. There was, however, a significant reduction in total flow during hypoxia in control animals (from 7·3 ±0·9 (10) to 5·2 ±0·5 (10) ml min⁻¹ kg⁻¹). The bradycardia and increased stroke flow during hypoxia were both abolished by atropinization and cardiac vagotomy (see Figs. 4 and 5).

Eight dogfish were first exposed to gradual hypoxia, allowed to recover, then injected with Propranolol and exposed for a second time to hypoxia. Hypoxia caused a significant bradycardia in control fish which was only partially compensated for by a rise in stroke flow. Total flow was significantly reduced to 78% of its normoxic value (from 8·2 + 0·9 (8) ml min⁻¹ kg⁻¹ to 6·4 ± 0·8 (8) ml ⁻¹ kg⁻¹). Following injection of Propranolol exposure to hypoxia still resulted in a significant bradycardia. This reduction in heart rate was not, however, accompanied by a rise in stroke flow (see Fig. 6). Total blood flow was, therefore, significantly reduced to 64% of its normoxic level (from 7 ± 1·1(8) down to 4·5± 0·7 (8) ml min⁻¹ kg⁻¹).

Butler & Taylor (1975) reported the absence of a significant bradycardia during gradual hypoxia down to a of 40 mmHg in dogfish acclimated to 7 °C. Stroke volume obtained indirectly via the Fick equation increased during hypoxia, resulting in an apparent increase in cardiac output. This result was checked for the present experimental conditions by measuring stroke flow directly. There was a significant bradycardia from a normoxic heart rate of 19·9 ± 1·1 down to 10·8 ± 1·2 beats min⁻¹ at a of 25 mmHg accompanied by a reduction in stroke flow which was not, however, statistically significant. The net result was that total blood flow was significantly reduced during gradual hypoxia, from 11 down to 6 ml min⁻¹ kg⁻¹, at a of 50 mmHg.

In two previous publications (Butler & Taylor, 1971, 1975) it was established that the heart of the dogfish operated under vagal tone in normoxic sea water at 12 °C and that it exhibited a bradycardia with an associated increase in cardiac stroke volume during hypoxia the extent of which varied with acclimation temperature. The present investigation examined in greater detail the role of the two pairs of branches of the cardiac vagus in the control of these responses, the relationship between heart rate and cardiac stroke volume, and the physiological role of the reflex bradycardia during hypoxia.

The means of all the measured variables in normoxia were similar to those previously reported for S. canicula by Butler & Taylor (1975). The injection of atropine caused an increase in mean heart rate in control fish acclimated to 12 °C (c.f. Butler & Taylor, 1971), 15 °C and 17 °C, indicating that the heart was operating under continuous, tonic vagal inhibition. Normoxic heart rate was, however, unaffected by atropinization at 7 °C. Several other authors have identified a vagal tone governing the rate of beating of the fish heart (e.g. Lutz, 1930; Izquierdo, 1930; Skramlik, 1935; Butler & Taylor, 1971) though the intensity of this tone is thought to vary between species (Stevens & Randall, 1967) and other authors have failed to demonstrate its presence in a range of fish species (Kulaev, 1957; Rodionov, 1959; Randall, 19^1 Randall & Smith, 1967). The present results indicate that the dogfish heart only operates under vagal tone at higher temperatures and that the intensity of the inhibition increases with temperature in a way which suggests that it may counteract the potential increase in heart rate with increasing body temperature (see Table 1). It is possible, therefore, that fish which have previously been described as either possessing or not possessing vagal tone on the heart were merely being studied either above or below a critical temperature at which the tonic influence of the vagus became apparent. The increasing intensity of vagal tone with temperature may relate to increased efferent activity in the branches of the cardiac vagus and/or to an enhanced sensitivity of the heart to acetylcholine as occurs in amphibians (e.g. Harri & Tirri, 1974). This problem is currently being investigated.

Total cardiac vagotomy caused an increase in heart rate which was very similar to the tachycardia caused by atropinization of control fish, and subsequent injection of atropine into vagotomized fish had no further effect on heart rate. These results indicate that atropinization was effective in removing the influence of the cardiac vagus. A similar tachycardia following cardiac nerve section in the dogfish was described by Lutz (1930). Priede (1974) described distinctly different effects of cardiac vagotomy in the rainbow trout at different acclimation temperatures. Heart rate increased following vagotomy at 6·5 °C but was unaffected at 15 °C. This is in sharp contrast to the present results where vagal tone on the heart increased with acclimation temperature. It is possible, however, that the chronotopic effects described by Priede (1974) are due to the variable influence of adrenergic fibres entering the teleostean heart (Gannon & Burnstock, 1969) or to circulating catecholamines (Bennion, reported by Stevens et al. 1972).

Selective sectioning of the branches of the cardiac vagus indicated that vagal tone is exerted predominantly via the branchial cardiac nerves, with the visceral cardiac nerves having a minor influence on the heart rate. In the series of experiments where the animals were allowed to recover for 2–3 days from the effects of cardiac nerve section, the visceral cardiac vagotomized fish had heart rates which were not significantly different from control fish, indicating that the tonic control exerted by the branchial cardiac nerves may increase to compensate for the removal of the activity in the visceral cardiac branches.

It is of course possible that part of the increase in heart rate resulting from vagotomy may be a consequence of deafferentating the heart. Laurent (1962) demonstrated an extensive afferent innervation of the teleost heart. The results of electrical stimulation of the central cut ends of the visceral cardiac and branchial cardiac branches of the vagus in the dogfish (Short et al.1977) have indicated that the relative afferent roles of the two pairs of cardiac nerves are roughly equivalent. Their unequal influence over heart rate appears, therefore, to relate to their efferent roles.

The increase in the normoxic heart rate following atropinization and cardiac vagotomy in control and sham-operated fish were both associated with a decrease in measured stroke flow in the aorta. This change compensated for the tachycardia and total blood flow was unchanged. This observation indicates that in terms of this response the dogfish under our experimental conditions responds to induced changes in heart rate according to Starling’s Law of the heart (Starling, 1915). Injection of the adrenergic β-receptor antagonist Propanolol indicated that there was no tonic adrenergic influence on the heart.

The responses to both rapid and gradual hypoxia were similar to those previously described by Butler & Taylor (1971, 1975). Of the measured variables only heart rate, stroke flow and showed significant changes during hypoxia. The bradycardia during hypoxia varied with temperature (c.f. Butler & Taylor, 1975). Atropine completely abolished the bradycardia during hypoxia at 12 °C. At 17 °C atropinization and total vagotomy were equally effective in abolishing the bradycardia. These results agree with previous work on both teleosts (Randall, 1966; Randall & Smith, 1967) and elasmobranchs (Satchell, 1961; Butler & Taylor, 1971).

The present results indicate that the branchial cardiac vagus was primarily responsible for the nervously mediated bradycardia during hypoxia at high temperatures. Visceral cardiac vagotomized fish showed a response to rapid hypoxia at 17 °C which was similar to normal fish, whereas branchial cardiac vagotomized fish showed a much reduced and modified response. This inequality between the two pairs of cardiac nerves resembles their different roles in the maintenance of vagal tone, reported above, and is reflected in the different degrees of cardiac inhibition resulting from electrical stimulation of their peripheral cut ends (Short et al.1977).

The potential reduction in blood flow, consequent upon the bradycardia during hypoxia, may be completely offset by an increase in stroke volume (Butler & Taylor, 1975). In the present investigation stroke flow increased during hypoxia in normal animals. This increase was, however, insufficient to compensate completely for the bradycardia induced in animals acclimated to 15 °C, and cardiac output decreased significantly during hypoxia in control animals. Short et al. (1977) found that relatively small reductions in heart rate (28%), brought about by electrically stimulating the cardiac vagi, were completely compensated for by a rise in stroke flow (33 %). Larger reductions in heart rate (52%) were, however, only partially compensated for by a rise in stroke flow of 38%, and total flow decreased significantly. This observation, plus the different experimental regime used by Butler & Taylor (1975), may explain the different degrees of compensation for a reduction in heart rate during hypoxia obtained in these two investigations.

Following the injection of the adrenergic β-receptor blocker Propranolol, stroke flow is unaffected by hypoxia, despite a marked bradycardia, resulting in a large decrease in calculated cardiac output. This result confirms the preliminary observation by Butler & Taylor (1975), who postulated that this may represent evidence for adrenergic control of the elasmobranch heart by circulating catecholamines, as described by Gannon et al. (1972). It has previously been shown that the blood levels of circulating catecholamines increase during asphyxie stress in elasmobranchs (Mazeaud, 1969). The possibility of independent adrenergic control of the inotropic responses of the heart to hypoxia is, however, contradicted by the effects of atropinization and total vagotomy. Both of these treatments abolished the bradycardia observed in control fish during gradual hypoxia and also completely abolished the associated rise in stroke flow (see Fig. 5), even when was reduced to very low levels (see Fig. 4). It seems possible, therefore, that the observed cardiac response to hypoxia in the dogfish, under the present experimental conditions, is entirely controlled by the vagus nerves and that the relationship between the chronotropic and inotropic responses follows Starling’s Law of the heart.

Direct measurement of blood flow in the ventral aorta during progressive hypoxia at 7 °C demonstrated that a significant bradycardia developed at low values of when that stroke flow was unaffected by hypoxia, and that there was a resultant drop in total flow. This result contradicts the observation of Butler & Taylor (1975) who obtained no significant reduction in heart rate during hypoxia, which was accompanied by an increase in derived cardiac output. In the present study when was reduced to 25 mmHg whilst Butler & Taylor (1975) exposed their fish to a minimum when of 39 mmHg. The critical when of 25 mmHg when mean heart rate was first significantly reduced from its normoxic level at 7 °C in this experimental series appears to correlate with the progressively higher values established by Butler & Taylor (1975) of 60 mmHg at 12 °C and 90 mmHg at 17 °C. Their inability to record a significant reduction in heart rate during hypoxia in fish acclimated to 7 °C may, therefore, be due to their having failed to reach the critical when for the change, or possibly to the longer (16 h) period of recovery from operations which they allowed their fish.

The physiological role of the bradycardia during hypoxia in gill-breathers has long been a subject for discussion (Hughes & Shelton, 1962; Heath, 1964). Satchell (1961), assuming that blood flow was significantly reduced, proposed that it may serve to maintain at as high a level as possible, presumably because it increased diffusion time in the gills for oxygen loading by the blood. In the present investigation atropinization had no effect on hypoxia at 7 °C (see Table 1). In this initial series of experiments there was no bradycardia associated with hypoxia at the point when the blood samples were taken. At 12 and 17 °C, however, there was a progressive reduction in hypoxic consequent upon atropinization which may relate to the progressively more intense bradycardia during hypoxia observed with increasing temperature (Butler & Taylor, 1975). The reduction in hypoxic at higher temperatures, following atropinization, would seem, therefore, to relate specifically to the abolition of the bradycardia during hypoxia. This is borne out by the effects of vagotomy. The hypoxic values measured from branchial cardiac vagotomized and totally cardiac vagotomized fish at 17 °C were significantly lower than those measured from control and sham-operated fish (see Fig. 3), and no further reduction in occurred following atropinization. It would seem that the nervously mediated bradycardia during hypoxia in 5. canicula acclimated to 17 °C serves to maintain the hypoxic at a level higher than that achieved when heart rate remains unchanged. Examination of an unpublished, in vitro, oxygen equilibrium curve for dogfish blood at 17 °C derived from the data published by Butler & Taylor (1975) reveals that the difference in hypoxic values in control and vagotomized animals are equivalent to a change from 20% to 8% saturation of the blood with oxygen. Also, the hypoxic of control fish at 17 °C is equivalent to approximately 2’5 × the hypoxic ⁠, whereas following vagotomy it is only ⁠, which may severely limit the amount of oxygen delivered to the tissues.

As well as the possibility that the bradycardia during hypoxia may in some way increase the diffusion time or capacity at the respiratory exchange surface, thus increasing oxygen transfer, there is the possibility that the reduction in the pumping activity of the heart significantly reduces the animal’s overall requirement for oxygen. Heath (1964) calculated that cardiac function accounts for approximately 26% of the standard oxygen consumption of the black grouper Mycteropera bonaci, though Garey (1970) calculated the proportion as 5% in the carp Cyprinus carpió. When the apparent cardiac stroke work and apparent power output of the heart were derived from the pressure drop across the heart and measured values for stroke flow (see Randall, 1970; Short, 1976) it appeared that apparent cardiac stroke work increased slightly (from 2·9 × 10⁻³ to 3·0 × 10³ J) in control fish acclimated to 15 °C, due to the measured increase in stroke flow. When, however, the equivalent values for apparent power output of the heart were derived from the respective heart rates there was a decrease from 5·7 × 10⁻⁴ to 3·4 × 10⁻⁴ W during hypoxia, as a consequence of the induced bradycardia. Although these values are far from a direct measure of the work performed by the heart they may indicate that the heart works less hard during the bradycardia induced by hypoxia.

The present investigation describes the responses to induced hypoxia of dogfish only recently recovered (3 h) from general anaesthesia and the implantation of various monitoring devices. Under these circumstances they show a clear reflex bradycardia during hypoxia, the extent of which varies with acclimation temperature. During normoxia the heart operates under continuous vagal tone which also varies with acclimation temperature. The changes in heart rate are to some extent compensated for by reciprocal changes in stroke flow in the ventral aorta, indicating that the system obeys Starling’s Law of the heart. The majority of the influence of the vagus nerve on heart rate appears to be mediated via the branchial cardiac vagi. The marked bradycardia shown by these fish during hypoxia, at high temperatures, may serve to maintain the hypoxic evel and also perhaps to reduce cardiac work, thus decreasing the animal’s demand for oxygen. Some information on the control of respiration and circulation in the dogfish, which may be useful in a developing study of the nervous control of the two systems has therefore been obtained.

This work was supported by grants from the Science Research Council. The authors wish to thank Professor E. W. Knight-Jones and the staff of the Zoology Department, University College of Swansea, in particular Mr Paul Llewellyn, for their co-operation. Also, we are grateful to the staff of the specimen department of the Marine Biological Association Laboratories in Plymouth for providing us with healthy dogfish.