Cardiac output has been measured directly, and calculated by the Fick method, during normoxia and hypoxia in six artificially perfused dogfish (Scyliorhinus canicula) in an attempt to estimate the accuracy of this method in fish. The construction and operation of a simple extra-corporeal cardiac bypass pump is described. This pump closely mimics the flow pulse profiles of the fish’s own heart and allows complete control of both cardiac stroke volume and systolic and diastolic periods.

During normoxia there was no significant difference between directly measured and calculated values for cardiac output. However, some shunting of blood past the respiratory surface of the gills may have been obscured by cutaneous oxygen uptake. In response to hypoxia there is either a decrease in the amount of blood being shunted past the respiratory surface of the gills and/or an increase in cutaneous oxygen uptake such that the Fick calculated value for cardiac output is on average 38% greater than the measured, value. It is proposed that the increase in the levels of circulating catecholamines that is reported to occur in response to hypoxia in this species may play an important role in the observed response to hypoxia. The results are discussed in terms of their implications for the calculation of cardiac output by the Fick principle in fish.

In studies of respiratory function and blood flow in fish, cardiac output is frequently calculated by use of the Fick principle (Randall, Holeton & Stevens, 1967; Butler & Taylor, 1975; Kiceniuk & Jones, 1977; Short, Taylor & Butler, 1979), where cardiac output equals oxygen consumption divided by the difference between the oxygen content of venous blood entering the gills, and the oxygen content of arterial blood leaving the gills. However, the use of the Fick principle can only be an accurate method for estimating the rate of blood flow from the heart if it is assumed that all the oxygen consumed by the fish is taken up across the gills and that all cardiac output flows across the gills and enters the systemic circulation via the dorsal aorta. This method for calculating cardiac output in fish has been borrowed from similar studies conducted on mammals in which the analogous assumptions are valid, but in fish this situation is more complex.

The importance of cutaneous gas exchange in many amphibians (Whitford & Hutchinson, 1965; Piiper & Gatz, 1974) and some reptiles (Belkin, 1968; Graham, 1974) has been appreciated for some time. However, in many studies of gas exchange in fish, the contribution that cutaneous oxygen uptake may make to the total oxygen consumption appears to be given little consideration. In eels at 11 °C, cutaneous oxygen uptake is reported to account for 35 % of the total oxygen consumed, while in trout the proportion is 13% (Kirsch & Nonnotte, 1977). In the antarctic ice fish Chaenocephatus aceratus cutaneous oxygen uptake has been estimated to contribute as much as 40% to the total oxygen consumption (Hemmingsen & Douglas, 1970). In a recent study, cutaneous oxygen uptake in the dogfish Scyliorhinus canicula was reported to contribute about 20% to the total oxygen consumption at 13 °C (Nonnotte & Kirsch, 1978). Both cutaneous oxygen uptake and oxygen consumption by the gill tissue (Johansen & Pettersson, 1981) will result in the total oxygen consumption being greater than the amount of oxygen being taken up by the gills and transferred to the blood, and consequently the calculated value for cardiac output will be larger than the actual cardiac output.

Recent studies of the gill vascular anatomy in many species of fish have revealed a complex vascular architecture, with both ‘respiratory’ and ‘non-respiratory’ blood pathways (Steen & Kruysse, 1964; Richards & Fromm, 1969; Vogel, Vogel & Kremers, 1973; Laurent & Dunel, 1976; Vogel, Vogel & Pfautsch, 1976; Dunel & Laurent, 1977, 1980; Cooke, 1980; Cooke & Campbell, 1980; Olson & Kent, 1980: J. D. Metcalfe & P. J. Butler, in preparation) and in many species it appears that some of the blood entering the gills from the heart may be diverted away from the gas exchange surface of the gills and return directly to the heart via the venous circulation of the gills and head. Any diversion of cardiac output within the gills in this manner will cause the calculated value for cardiac output to be smaller than the actual value.

In a recent study of the eel, Hughes et al. (1981) report that as much as 30% of the total cardiac output is shunted away from the respiratory surface of the gills and is not involved in gas exchange, this blood presumably returns directly to the heart via the venous circulation. In addition, these authors report that adrenaline (2 μg kg-1) reduced the proportion of cardiac output that is shunted away from the respiratory surface to about 6 % indicating a greater blood flow to the gas exchange surface, and also reducing the difference between the calculated and actual values for cardiac output.

In the present study, the Fick method for the calculation of cardiac output in the dogfish has been evaluated by comparing simultaneously calculated and directly measured values. Hypoxia causes an increase in the levels of circulating adrenaline and noradrenaline in Scyliorhinus canicula (Butler et al. 1978; Butler, Taylor & Davison, 1979) and since these hormones may affect the proportion of cardiac output which enters the systemic circulation, the comparison has been made both during normoxia and hypoxia.

In the dogfish, the anatomical arrangement of the afferent branchial arteries with.respect to the heart, render it impossible to measure total cardiac output directly ‘with a cannulating electromagnetic flow measuring device, without entering the pericardium. The latter procedure is undesirable since the integrity of the rigid pericardium in elasmobranchs is reported to be essential for normal cardiac function (Hanson, 1967). For these reasons, circulation has been maintained in the dogfish in the present study by means of an extra-corporeal cardiac bypass which has been developed for this purpose.

A number of cardio-respiratory studies in intact fish have employed an extra corporeal cardiac bypass (Saunders & Sutterlin, 1971; Kent & Pierce, 1978 ; Opdyke, Holcombe & Wild, 1979; Hughes et al. 1981). Any such system should mimic as closely as possible the flow pulse profiles of the fish’s own heart (cf. Bacon et al. 1976) and should allow independent control of stroke volume, and systolic and diastolic periods during operation as well as maintaining the blood at the experimental temperature. These criteria should be fulfilled with the minimum of cellular damage to the blood. The construction of such an extra-corporeal cardiac bypass is described in in the present report.

Construction of an extra-corporeal cardiac bypass

The blood pump (Fig. 1) is a modification of the design by Daly, Ead & Scott (1978). Pulsatile flow is induced by the sinusoidal compression of a portion of an 18 cm loop of silicon rubber tube (I.D. 6 mm, O.D. 9 nuns Portex Ltd.) by a triangular compression plate which is moved, via a Perspex drive shaft, by an off-centre drive cam mounted on the shaft of a high torque (3·5 lb in) electric motor (Klaxon Ltd.). The power to the motor is supplied from the control unit of a commercial peristaltic pump (Watson-Marlow, MHRE 22). Silicon rubber tube has a high coefficient of restitution and good biocompatability (M. de B. Daly, personal communication) and was used for these reasons. Commercial plate valves (Tudor Accessories Ltd.) at both ends of the silicon rubber tube maintain a unidirectional flow. In each valve the plate is loaded by a small stainless-steel spring which effectively prevents any back flow due to its short response time in closing. The valves are retained in a Perspex valve block which is mounted on a Perspex anvil such that the loop of silicon rubber tube lies across the upper surface of the anvil which provides support during compression. The anvil, with the valve/tube assembly, is mounted within the Perspex water bath of the pump on a 27 cm length of o BA stainless steel studding connected to a handle mounted on the outside of the water bath. Rotation of this handle allows the anvil and the valve/tube assembly to be moved along a track on the base of the water bath and so the length of the silicon rubber tube under the compression plate can be varied. In this way stroke volume of the pump may be altered independently of any other variable during operation. The drive shaft is mounted on the side of the water bath by two Perspex yokes which are pivoted at either end on brass/steel bearings. A round Perspex cap is mounted on the top of the drive shaft on a 1 cm long o. BA thread. Rotation of this cap allows alteration of the length of the drive shaft, and thus the proportion of the drive cam which makes contact with the drive shaft can be controlled. Wear on the cap of the drive shaft by rotation of the cam is reduced by a brass roller-bearing, mounted in a Perspex yoke, situated between the cam and the cap, which is also mounted on the side of the water bath.

Fig. 1.

The extra-corporeal cardiac bypass pump: (A) Front view. (B) Top view of anvil and valve block assembly, showing position of compression plate. (C) Side view. A, Anvil ; BT, bubble trap; C, cap; CP, compression plate; DC, drive cam; DS, drive shaft; H, handle; MS, micro switch; M, motor; RB, roller bearing; ST, silicon rubber tube; S, stainless steel studding; T, track; TC, tripping cam; V, valve; VB, valve block; WB, water bath; Y, yokes.

Fig. 1.

The extra-corporeal cardiac bypass pump: (A) Front view. (B) Top view of anvil and valve block assembly, showing position of compression plate. (C) Side view. A, Anvil ; BT, bubble trap; C, cap; CP, compression plate; DC, drive cam; DS, drive shaft; H, handle; MS, micro switch; M, motor; RB, roller bearing; ST, silicon rubber tube; S, stainless steel studding; T, track; TC, tripping cam; V, valve; VB, valve block; WB, water bath; Y, yokes.

Systolic and diastolic periods are independently controlled by means of a micro-switch which is activated by a tripping cam mounted on the shaft of the motor behind the drive cam. The micro-switch and tripping cam are set so that the switch is open during the down stroke of the drive shaft (systolic period) but closed at all other times (diastolic period). In this way the power supply to the motor is switched between two 10 turn potentiometers which control the motor speed. This allows the motor speed, and thus the period during systole and diastole, to be controlled independently of each other during operation of the pump.

Control of the pump in this way allows stroke volume to be continuously variable between o and about 4·0 ml. Systolic and diastolic periods are continuously variable between 0·2 and 3·0 s giving an overall pulse frequency range of 10–150 min-1. The values of these variables span well outside those values measured in intact dogfish of the size range used in the present study (c.f. Short et al. 1979).

Blood flowing to and from the pump passes through small glass bubble traps (Fig. 1) which contain a small, variable, volume of air (about 3–5 ml), which acts as a ‘Wind-kessel’. On the inflow side of the pump, the ‘Windkessel’ effect of the bubble trap reduces the violence of the diastolic filling which was found to be important in maintaining adequate venous return to the pump. On the outflow side of the pump, the ‘Windkessel’ effect of the bubble trap damps out the high-frequency oscillations observed in flow pulses at very low frequencies (less than 15 min-1) which appear to be due to valve ‘flutter’. These damping chambers do not affect the overall pulse duration. The temperature of the blood flowing through the pump is controlled by passing water at the desired temperature through the water bath surrounding the valve/tube assembly.

Preparation of animals

In the present study cardiac output has been simultaneously calculated via the Fick principle, and measured directly, during normoxia and induced environmental hypoxia in six lesser spotted dogfish (Scyliorhinus canicula) of either sex, the mass of which ranged from 0·603 to 0·897 kg. These were obtained from the Plymouth Laboratories of the Marine Biological Association of the U.K., and transported in oxygenated sea water to the aquaria in Birmingham where they were held in aerated, recirculating sea water maintained at 15 ± 1 °C for at least two weeks prior to any experiments. Fish were fed periodically on either whitebait or sprats obtained from a local fishmonger. All experiments were performed at the above temperature, and each fish was starved for 4 days prior to any experiment. Having been washed with strongly heparinized (2000 units i.u. ml-1 sodium heparin, Weddel) elasmobranch Ringer’s solution (Capra & Satchell, 1977 a), the perfusion circuit (the pump and associated cannulae) was primed with 30–40 ml of heparinized (20 units i.u. sodium heparin ml-1, Weddel) dogfish blood obtained from a donor fish. This blood recirculated around the perfusion circuit prior to the cannulation of the experimental fish.

The experimental fish was anaesthetized in sea water containing 0·04 g I-1 tricaine methanesulfonate (MS 222, Sigma Chemical Co.) and placed ventral side up on an operating table in a constant temperature room maintained at 15 °C. The gills were irrigated with recirculating, filtered, aerated sea water containing anaesthetic (as above). The caudal artery was exposed via a 15 mm longitudinal, ventral incision just posterior to the anal fin, and cannulated with a 40 cm length of polythene tubing (o.D. 1·0 mm, Portex) filled with heparinized elasmobranch Ringer’s. This cannula allowed the measurement of dorsal aortic blood pressure and cardiac pump rate, and the sampling of post branchial blood for the measurement of arterial blood oxygen contents and tensions. The caudal vein was plugged at this point with a 1 cm length of heat-sealed polythene tubing (as above). The wound was sutured closed and sealed with a patch of household rubber glove attached with cyanoacrylate adhesive (R.S. Components Ltd.) to prevent the loss of blood. Heparin (2000 units i.u. sodium heparin kg-1, as above) was injected via this cannula into the fish’s blood stream and allowed to circulate for 5-10 min before continuing the surgical procedure.

The pericardium was exposed and opened via a 20 mm ventral incision. The junction between the ventricle and the conus arteriosus was ligated to prevent blood loss and the ventral aorta was cannulated via the conus arteriosus with a 100 cm length of blood-filled PVC tubing (o.D. 2-8 mm, Portex) which was connected to the outflow from the pump via one of the bubble traps. This bubble trap was fitted with a short cannula to allow sampling of venous blood leaving the pump (Fig. 2). Care was taken to prevent any air from entering the circulation during this part of the procedure. Once this cannula had been tied into place, the sinus venosus was cannulated with a similar length of PVC tubing which was connected to the inflow to the pump via the second bubble trap. The cannula placed in the sinus venosus was tipped with a small perforated cage (7 × 2·8 mm O.D., curved through 45°) manufactured from ‘Araldite’, the tip of which lay in the opening of one of the Cuverian ducts. This cage prevented the collapse of the sinus venosus and allowed venous blood to be drawn into the pump under negative pressure. During the cardiac bypass procedure, blood flow was arrested for about 15-20 min. However, this did not appear adversely to affect the fish. Once the cannulae had been tied into place, perfusion was commenced at a low pulse rate (about 8 beats min-1) with a stroke volume of rond kg-1 (Short et al. 1979). The pericardio-peritoneal canal was plugged with a small glass plug (10×4 mm diam.) to prevent any blood loss subsequently leaking into the peritoneum from the pericardium. The wound was sutured closed and sealed as previously described.

Fig. 2.

The experimental arrangement used in obtaining the physiological measurements from dogfish artificially perfused with blood by the extra-corporeal cardiac bypass pump. D, Dogfish; A, dorsal aortic cannula; Q, electromagnetic flow meter; E, experimental tank; P, extra-corporeal pump; F, filter; G, gas exchange columnsPinc,o2, inflow to respirometer; Pex,o2 outlet from respirometer; ▸ direction of sea water flow; ▸, direction of blood flow; R, respirometer; T, sea Water flow meter; V, venous canula; W, water bath.

Fig. 2.

The experimental arrangement used in obtaining the physiological measurements from dogfish artificially perfused with blood by the extra-corporeal cardiac bypass pump. D, Dogfish; A, dorsal aortic cannula; Q, electromagnetic flow meter; E, experimental tank; P, extra-corporeal pump; F, filter; G, gas exchange columnsPinc,o2, inflow to respirometer; Pex,o2 outlet from respirometer; ▸ direction of sea water flow; ▸, direction of blood flow; R, respirometer; T, sea Water flow meter; V, venous canula; W, water bath.

The fish was placed in a Perspex respirometer (capacity 5 1) and restrained by a clamp at the base of its tail. The cannulae passed through bungs sealed into the lid of the respirometer. Once sealed, the respirometer was placed in a black Perspex experimental tank containing about 40 1 filtered, aerated, recirculating sea water at 15 ± I °C which covered the respirometer. The experimental tank was covered with polystyrene foam so that the fish received the minimum of visual disturbance. The respirometer was supplied with sea water drawn from the experimental tank via a pair of gas exchange columns placed in series, at a flow rate of approximately 0·75 1 min-1, measured with a flow meter (Rotameter 1100 sea water: GEC-Elliot) (Fig. 2). The paired gas exchange columns allowed accurate control of the oxygen tension of the water flowing into the respirometer during hypoxia independently of the oxygen tension of the water being drawn from the experimental tank.

The dorsal aortic cannula was connected to a pressure transducer (s.E. Labs: S.E.M. 4·86) for the measurement of dorsal aortic blood pressure, and its output was displayed on a four-channel rectilinear pen recorder (Ormed Ltd.). Blood flow from the cardiac pump was measured with a cannulating electromagnetic flow probe (Biotronex: 1·5 mm I.D.) placed in the ventral aortic cannula. The flow probe was connected to a Biotronex pulsed logic flow meter (BL-610) and the output displayed on the pen recorder (as above). The flow probe was calibrated with the fish’s own blood at the end of each experiment. Oxygen tensions of blood and water samples were measured with a blood gas analyser (Radiometer PHM 71) having its oxygen electrode housed in a glass cuvette maintained at the experimental temperature. The oxygen content of blood samples was measured with a Lex-O2-Con total oxygen analyser (Lexington Instruments) which was calibrated with oxygen-saturated, distilled water at o °C so as to be accurate over the range of low oxygen contents encountered in blood samples ( < 2·3 m mol I-1). Haematocrit was measured with a microhaematocrit centrifuge (Hawksley) from venous blood samples.

Once the fish had been placed in the experimental tank, the cardiac pump rate was gradually raised to about 30 beats min-1 (Short et al. 1979) over a period of about 5 min. In a number of preliminary experiments it was found that adequate venous return to the pump could only be maintained by increasing the blood volume of the fish by about 25 ml kg-1. Consequently all fish received extra blood from a donor fish in approximately this proportion.

Fish were allowed 5 h to recover from the anaesthetic in normoxic water ( = about 21 kPa). At the end of the period measurements of normoxic cardiac pump rate and stroke volume were recorded over a 3 min period and directly measured cardiac output was calculated as stroke volume x cardiac pump rate. Mean dorsal aortic blood pressure was calculated as diastolic blood pressure + (systolic-diastolic blood pressure). The oxygen tensions of water incurrent and excurrent ( to the respirometer, and respirometer water flow, were measured1 and total oxygen consumption calculated by the method described by Short et al. (1979). Blood samples of between 0·5–·7 ml were drawn into glass syringes from the dorsal aorta (arterial blood) and from the bubble trap on the outflow side of the cardiac pump (venous blood). The oxygen contents (and ) and oxygen tensions (and ) of these samples were measured and cardiac output was calculated via the Fick principle ( cal.) by the method described by Short et al. (979).

The oxygen tension of the water incurrent to the respirometer was then reduced to about 8·6 kPa over a period of 30 min by passing nitrogen at an appropriate rate through the second of the two gas exchange columns. This oxygen tension was maintained for the rest of the experimental period. As the water oxygen tension was reduced, the reflex bradycardia observed in intact dogfish in response to environmental hypoxia (Butler & Taylor, 1975; Short et al. 1979) was simulated with the cardiac pump; the pulse rate was halved and the stroke volume doubled thus maintaining cardiac output at the normoxic value. Tests showed that at the water flow rates used (0·75 1 min-1) the oxygen tension across the respirometer equilibrated after about 20–30 min. Accordingly animals were allowed 1 h to adjust to the reduced oxygen tension and the above variables were again measured and cardiac output was calculated and measured directly.

In four experiments the oxygen consumption of any organism living upon the skin of the fish was estimated. At the end of the experiment the fish was killed in situ by injecting either sodium cyanide or sodium azide (2 mg kg-1) into the blood stream of the fish. This was allowed to circulate for a few min and then the cardiac pump was turned off. One h after death, total oxygen consumption in normoxic water was measured as previously described. After such metabolic poisoning, there was no detectable oxygen consumption. It appears that any organisms that live on the skin of the dogfish make no measurable contribution to the calculated value for total oxygen consumption.

In the present report, all mean values are expressed ± S.E.M. The difference between means has been compared by using the paired t-test (Bailey, 1959). The term ‘significant* refers to the 95 % level of confidence (P < 0·05), unless otherwise stated.

The operation of the extra-corporeal cardiac bypass as reported maintained the viability of the fish for the entire experimental period in all cases, and in one preliminary experiment an individual fish remained alive for 21 h, after which time the experiment was terminated. Fig. 3 shows that flow pulse profiles in artificially perfused fish were similar to those obtained from intact dogfish (Short, 1976) although the falling phase of the flow pulse is somewhat longer in perfused fish. Though some haemolysis of the blood might have been expected due to the operation of the cardiac pump, none was apparent. In both haematocrit samples, and in blood which had been allowed to settle overnight after the perfusion experiments, the plasma was always clear with no obvious red colouration.

Fig. 3.

Comparison of pulse flow waves from an intact dogfish (a) (from Short, 1976) and from a dogfish perfused with blood using the extra-corporeal cardiac bypass pump (b). The time marker (t) indicates one second intervals.

Fig. 3.

Comparison of pulse flow waves from an intact dogfish (a) (from Short, 1976) and from a dogfish perfused with blood using the extra-corporeal cardiac bypass pump (b). The time marker (t) indicates one second intervals.

Values for cardiac output measured directly and calculated by the Fick technique

During Normoxia and Hypoxia

The mean ± S.E.M. of the mean of the variables measured in intact, artifically perfused dogfish during both normoxia and hypoxia are presented in Table 1. The individual and mean differences between cal. and mea. (presented as the ratio between these two parameters) during normoxia and in response to hypoxia for the 6 fish are illustrated in Fig. 4.

Table 1.

The mean ( ± s.E. of mean) values of the measured variables in unanaesthetised dogfish artificially perfused with blood, at rest during normoxia and hypoxia

The mean ( ± s.E. of mean) values of the measured variables in unanaesthetised dogfish artificially perfused with blood, at rest during normoxia and hypoxia
The mean ( ± s.E. of mean) values of the measured variables in unanaesthetised dogfish artificially perfused with blood, at rest during normoxia and hypoxia
Fig. 4.

A graphical representation of the ratio between calculated cardiac output V˙b and directly measured cardiac output V˙b in six artificially perfused dogfish during normoxia (•) and in response to hypoxia (Δ). Mean values, ±s.B.M. are also shown.

Fig. 4.

A graphical representation of the ratio between calculated cardiac output V˙b and directly measured cardiac output V˙b in six artificially perfused dogfish during normoxia (•) and in response to hypoxia (Δ). Mean values, ±s.B.M. are also shown.

In the present study during normoxia, the values for arterial and venous blood oxygen contents and tensions, and dorsal aortic blood pressure in artifically perfused dogfish are similar to those reported for intact dogfish at 15 °C (Short et al. 1979), although the present value for oxygen consumption is about 75% °f the value reported by these authors. This appears to confirm the visual observation that dogfish tolerate artificial perfusion well, at least over the period of the present experiments (about 9 h).

In normoxic animals the calculated value for is on average 8% lower than the directly measured value, however, this difference is not significant. . values during normoxia cover a broad span from 0·69 to 1·24 (Fig. 4).

In response to hypoxia both aterial and venous blood oxygen contents and tensions decreased significantly and this was accompanied by an 18% decrease in oxygen consumption, however, this decrease was not significant. There was a small reduction in directly measured following adjustment of the cardiac pump during hypoxia, however this change was not significant. The changes in the measured variables in response to hypoxia in the present study are similar to those reported for this species by Short et al. (1979). in response to a similar reduction in environmental oxygen tension .

In contrast to normoxic fish, calculated is much (38%) higher than measured during hypoxia. This difference is significant at the 90% level (P < 0·1). . values during hypoxia ranged from o-88 to 2·09 (Fig. 4). Since blood flow cannot be greater than that actually measured, this difference may indicate that a large proportion of the total oxygen consumption is being taken up across the skin. In response to hypoxia the /. ratio increased in all six fish, the mean value increasing from 0·92 during normoxia to 1·38 in response to hypoxia, and this increase is significant at the 95% level of confidence (P < 0·05). This indicates a marked increase in observed cutaneous oxygen uptake during hypoxia (Fig. 4).

In the present study dogfish survived artificial perfusion well for periods of at least 9 h. Mean dorsal aortic blood pressure, and arterial and venous blood oxygen tensions in perfused fish were similar to those previously reported for this species (Short et al. 1979). This would indicate that normal gas exchange is maintained in the present study. Although a number of studies have reported artificial perfusion to be successful in a number of fish species (see Introduction for references), few appear to have succeeded in maintaining viability for as long as 9 h. This may be related to the ability of the extra-corporeal cardiac bypass reported in the present study to mimic closely the flow characteristics of the fish’s own heart.

During normoxia there was no significant difference between cal. and mea. which indicates that there was no shunting of blood past the respiratory surface of the gills despite the fact that such a shunt, via the vascular network of the interbranchial septum, has been shown to be anatomically possible in S. canicula (J. D. Metcalfe & P. J. Butler in preparation). However, the present study only reveals the minimum value for the shunting of blood past the respiratory surface during normoxia. If, in normoxic fish, some part of the total oxygen consumed is being taken up across the skin, then the actual proportion of cardiac output that is being shunted past the respiratory surface of the gills will be larger than that estimated in the present study since, by the method of calculating by the Fick technique, cutaneous oxygen uptake and the shunting of blood past the gas exchange surface of the gills mutually obscure each other. If the estimate of 20% for cutaneous oxygen uptake in this species made by Nonnotte & Kirsch (1978) is accurate, then the real proportion of cardiac output that is shunted past the gas exchange surface may be as high as 28 %. However, this value of 20% for cutaneous oxygen uptake made by Nonnotte & Kirsch (1978) may be unusually high due to their experimental method which would have destroyed the stable boundary layers of water next to the skin of the fish, causing oxygen uptake to increase.

The fact that normoxic oxygen consumption in the present study is lower than that previously reported for intact dogfish at 15 °C (Short et al. 1979), despite the fact that the arterio-venous oxygen content difference is similar, would suggest that cardiac output in the present study was lower than that measured by Short et al. (1979). However, the measured value for in the present study is similar to that calculated by the Fick technique by these authors. This indicates that the actual blood flow across the gas exchange surface of the gills in the present study may have been lower than in the earlier study.

In response to hypoxia oxygen consumption decreased by 18%. Although this reduction was not significant in the present study, similar but significant reductions in oxygen consumption have previously been reported in response to hypoxia both in this species (Butler & Taylor, 1975; Taylor, Short & Butler, 1977; Short et al. 1979) and other elasmobranchs (Piiper, Baumgarten & Meyer, 1970) at environmental temperatures above 7 °C. During hypoxia cutaneous oxygen uptake appears to increase and accounts for almost 40% of the total oxygen consumed. This dramatic increase in observed cutaneous oxygen uptake during hypoxia may be the result of either haemodynamic changes within the gills which reduce any blood shunt, thereby revealing the actual value for cutaneous oxygen uptake, or of a real increase in cutaneous oxygen uptake, or probably some combination of these two processes.

How such changes in the regional distribution of branchial blood flow and oxygen transfer across the skin are brought about in response to hypoxia are as yet unclear. However, they may be related to the increase in the levels of circulating catecholamines adrenaline and noradrenaline which have previously been reported to occur in response to hypoxia in S. canicula (Butler et al. 1978, 1979). These hormones have repeatedly been reported to enhance blood flow across the respiratory surface of the gills of both teleost and elasmobranch fish (Rankin & Maetz, 1971; Randall, Baumgarten & Malyuse, 1972; Bergman, Olson & Fromm, 1974; Girard & Payan, 1976; Dunel & Laurent, 1977; Hughes et al. 1980). The possible role of these humoral agents may be of particular importance in the dogfish since this species appears to lack any direct neural control of its gill blood vessels (J. D. Metcalfe & P. J. Butler, in preparation). Noradrenaline has also been reported to cause vasodilatation of the skin blood vessels in elasmobranchs (Capra & Stachell, 1977b) and this may serve to enhance oxygen transfer across the skin during hypoxia.

The present study does not reveal the absolute values for either the proportion of cardiac output that bypasses the gas exchange surface, or for the proportion of the total oxygen consumption which is taken up across the skin in 5. canicula. However, it does demonstrate that the calculation of any variable by the Fick technique which assumes that all oxygen is taken up across the gills, and that all cardiac output reaches the dorsal aorta, may not only be inaccurate, but may also not reveal the real changes in these variables that occur in response to hypoxia.

Financial support was provided by the Science and Engineering Research Council.

Bacon
,
D. R.
,
Daly
,
C. de B.
,
Daly
,
M. D. B.
&
Scott
,
R. W.
(
1976
).
Modified roller pump with improved haemodynamic characteristics and temperature control of perfusate
.
Laboratory Practice
25
,
464
466
.
Bailey
,
N. T. J.
(
1959
).
Statistical Methods in Biology
.
English Universities Press
.
London
.
Belkin
,
D. A.
(
1968
).
Aquatic respiration and underwater survival of fresh water turtle species
.
Retpir. Phytiol
.
4
,
1
14
.
Bergman
,
H. L.
,
Oleen
,
K. R.
&
Fromm
,
P. O.
(
1974
).
The effects of vasoactive agents on the functional surface area of the isolated perfused gills of the rainbow trout
.
J. comp. Physiol
.
94
,
267
286
.
Butler
,
P. J.
&
Taylor
,
E. W.
(
1975
).
The effects of progressive hypoxia on respiration in the dogfish (Scyliorltimu canicula L.) at different seasonal temperatures
.
J. exp. Biol
.
63
,
117
130
.
Butler
,
P. J.
,
Taylor
,
E. W.
,
Capra
,
M.
&
Davison
,
W.
(
1978
).
The effect of hypoxia on the levels of circulating catecholamines in the dogfish Scyliorhinus canicula L
.
J. comp. Phytiol. B
127
,
325
330
Butler
,
P. J.
,
Taylor
,
E. W.
&
Davison
,
W.
(
1979
).
The effect of long term, moderate hypoxia on acid-base balance, plasma catecholamine and anaerobic end products in the unrestrained dogfish Scyliorhinus camada
,
J. comp. Physiol. B
132
,
297
303
.
Capra
,
M. F.
&
Satchell
,
G. H.
(
1977a
).
Adrenergic and cholinergic responses of the isolated saline-perfused heart of the elasmobranch fish Squalus sucklyei
.
Comp, biochem. Physiol
.
8
,
56
65
.
Capra
,
M. F.
&
Satchell
,
G. H.
(
1977b
).
The differential haemodyanic response of the elasmobranch, Squalua acanthias to the naturally occurring catecholamines adrenaline and noradrenaline
.
Comp. Biochem. Physiol
.
58C
,
41
48
.
Cooke
,
I. R. C.
(
1980
).
Functional aspects of the morphology and vascular anatomy of the gills of the endeavour dogfish, Centrophorus scalpratus (McCulloch) (Elasmobranchii : Squalidae)
.
Zoomorphologie
94
,
167
183
.
Cooke
,
I. R. C.
&
Campbell
,
G.
(
1980
).
The vascular anatomy of the gills of the smooth toadfish, Torquiginer glaber (Teleostii: Tetraodontidae)
.
Zoomorphologie
94
,
151
166
.
Daly
,
M. DE B.
,
Ead
,
H. W.
&
Scott
,
R. W.
(
1978
).
A small pulsatile perfusion pump-blood equilibrator system
.
J. Physiol., Land
.
280
,
6
8p
.
Dunel
,
S.
&
Laurent
,
P.
(
1977
).
La vascularisation branchiale chez 1’Anguille: action de l’acétylcholine et de l’adrénaline sur la répartition d’une résine polymérisable dans les differénts compartiments vasculaires
.
C.r. hebd. Seanc. Acad. Sci. Paris
,
284
,
2011
2014
.
Dunel
,
S.
&
Laurent
,
P.
(
1980
).
Functional organisation of the gill vasculature in different classes of fish
.
In epithelial Transport in the Lower Vertebrates
(ed.
B.
Lahlou
), pp.
37
58
.
Cambridge
:
Cambridge University Press
.
Girard
,
J. P.
&
Payan
,
P.
(
1976
).
Effect of epinephrine on vascular space of gills and head of rainbow trout
.
Am. J. Physiol
.
230
,
1555
1560
.
Graham
,
J. B.
(
1974
).
Aquatic respiration in the sea snake Pelamis platurus
.
Respir. Physiol
.
21
,
1
7
.
Hanson
,
D.
(
1967
).
Cardiovascular dynamics and aspects of gas exchange in Chondrichthyes
,
Ph.D. Dissertation, University of Washington, Seattle, Washington (quoted by Satchell, G. H.) in Circulation in Fishes (1971)
,
Cambridge University Press, Cambridge
).
Hemmingsen
,
E. A.
&
Douglas
,
E. L.
(
1970
).
Respiratory characteristics of the haemoglobin-free fish Chaenocephalus aceratus
.
Comp. Biochem. Physiol
.
32
,
733
744
.
Hughes
,
G. M.
,
Peyraud
,
C.
,
Peyraud-Waitzeneggkr
,
M.
&
Soulier
,
P.
(
1981
).
Proportion of cardiac output concerned with gas exchange in the gills of the eel (A. anguilla)
.
J. Physiol., Lond
.
310
,
61
62P
.
Johansen
,
K.
&
Pettersson
,
K.
(
1981
).
Gill O2 consumption in a teleost fish (Gadus morhua)
.
Respir. Physiol
.
44
,
277
284
.
Kent
,
B.
&
Pierce
,
E. C.
(
1978
).
Cardiovascular responses to changes in blood gases in dogfish shark, Squalus acanthias
.
Comp. Biochem. Physiol
.
60C
,
37
44
.
Kiceniuk
,
J. W.
&
Jones
,
D. R.
(
1977
).
The oxygen transport system in trout (Salmo gairdneri) during sustained exercise
.
J. exp. Biol
.
69
,
247
260
.
Kirsch
,
R.
&
Nonnotte
,
G.
(
1977
).
Cutaneous respiration in three fresh water teleosts
.
Resp. Physiol
.
29
,
339
354
.
Laurent
,
P.
&
Dunel
,
S.
(
1976
).
Functional organisation of the teleost gill. I; Blood pathways
.
Acta Zool
.
57
,
189
209
.
Nonnotte
,
G.
&
Kirsch
,
R.
(
1978
).
La respiration cutanée chez un selacian (Scyliorhinus canicula)
.
C.r. hebd. Acad. Sd. Seanc. Paris
.
286
,
1597
1599
.
Olson
,
K. R.
&
Kent
,
B.
(
1980
).
The microvasculature of the elasmobranch gill
.
Cell Tissue
.
209
,
49
63
.
Opdyke
,
D. F.
,
Holcombe
,
R.
&
Wild
,
D. W.
(
1979
).
Blood flow resistance in Squalus acanthias
.
Comp. Biochem. Physiol
.
62A
,
711
717
.
Piiper
,
J.
,
Baumgarten
,
D.
&
Meyer
,
M.
(
1970
).
Effects of hypoxia upon respiration and circulation in the dogfish (Scyliorhinus stellaris)
.
Comp. Biochem. Physiol
.
36
,
513
520
.
Piiper
,
J.
&
Gatz
,
R. N.
(
1974
).
Exploitation des méchanismes des échangés gazeux chez une salamandre ne possédant ni poumons, ni branchies
.
J. Physiol., Paris
69
,
287
A.
Randall
,
D. J.
,
Holeton
,
G. F.
&
Stevens
,
E. D.
(
1967
).
The exchange of oxygen and carbondioxide across the gills of the rainbow trout
.
J. exp. Biol
.
46
,
339
348
.
Randall
,
D. J.
,
Baumgarten
,
D.
&
Mayluse
,
M.
(
1972
).
The relationship between gas and ion transfer across the gills of fish
.
Comp. Biochem. Physiol
.
41 A
,
629
637
.
Rankin
,
J. C.
&
Maetz
,
J.
(
1971
).
A perfused teleostean gill preparation: vascular actions of neurohypophysial hormones and catecholamines
.
J. Endocr
.
51
,
621
635
.
Richards
,
B. D.
&
Fromm
,
P. O.
(
1969
).
Patterns of blood flow through filaments and lamellae of isolated-perfused rainbow trout gills
.
Comp. Biochem. Physiol
.
29
,
1063
1070
.
Saunders
,
R. L.
&
Suttbrlin
,
A. M.
(
1971
).
Cardiac and respiratory responses to hypoxia in the sea raven, Hemitripterus americanus, and an investigation of possible control mechanisms
.
J. Fish. Res. Bd. Can
.
28
,
491
503
.
Short
,
S.
(
1976
).
Some aspects of cardioregulation in the dogfish (Scyliorhinus canicula)
.
Ph.D. thesis
.
University of Birmingham, U.K
.
Short
,
S.
,
Taylor
,
E. W.
&
Butler
,
P. J.
(
1979
).
The effectiveness of oxygen transfer during normoxia and hypoxia in the dogfish (Scyliorhinus canicula) before and after cardiac vagotomy
.
J. comp. Physiol. B
132
,
289
295
.
Steen
,
J. B.
&
Kruysse
,
A.
(
1964
).
The respiratory function of the teleostean gills
.
Comp. Biochem. Physiol
.
12
,
127
142
.
Taylor
,
E. W.
,
Short
,
S.
&
Butler
,
P. J.
(
1977
).
The role of the cardiac vagus in the response of the dogfish Scyliorhinui canicula to hypoxia
.
J. exp. Biol
,
70
,
57
75
.
Vogel
,
W.
,
Vogel
,
V.
&
Kremers
,
H.
(
1973
).
New aspects of the intrafilamental vascular system in the gills of an euryhaline teleost Tilapia moitambica
.
Z. Zellforich
.
144
,
573
583
.
Vogel
,
W.
,
Vogel
,
V.
&
Pfautsch
,
M.
(
1976
).
Arterio-venous anastomoses in rainbow trout gill filaments
.
Cell Tiitue Ret
.
167
,
373
385
.
Whitford
,
W. G.
&
Hutchison
,
V. H.
(
1965
).
Gas exchange in salamanders
.
Physiol. Zool
.
38
,
228
242