On The Respiration of Torpedo Marmorata

One problem met during investigations of fish respiration concerns their activity and its variation during experiments. Many workers have attempted to quantify such activity, and some ingenious apparatus has been developed in recent years, which enables studies to be made of fish when swimming at controlled speeds (Brett, 1964, 1972; Webb, 1971; Beamish, 1970). Another solution has been to study fish which spend a substantial amount of time in an inactive condition; among teleosts, flatfish such as plaice and other bottom-living forms such as Callionymus(Hughes & Umezawa, 1968b; Hughes & Knights, 1968) have provided information of general interest. Among elasmobranchs, bottom-living forms such as rays (Raia sp.) have also been studied but, although generally less active than dogfish, they are more active than Torpedo, which may remain stationary in an aquarium for several days. This paper describes observations on the respiration of this fish which developed from initial studies on the thornback ray (R. clavata). A particular interest in Torpedo from a physiological point of view is their response to hypoxia, as it was found that specimens remain inactive even at very low oxygen tensions and hence provide useful material for investigating the changes which occur during such extreme respiratory stresses.

Most studies were carried out at the Institut de Biologie marine, of the University of Bordeaux at Arcachon. Specimens of Torpedo marmorata Risso, ranging in body weight from 500 g to 5 kg, were collected by fishermen from the Gulf of Gascony. They were kept in the sea-water circulation of the laboratory and used at least 1 week following capture and were generally kept without feeding.

Oxygen consumption was measured using three methods :

Fish were kept in a ray-shaped respirometer and maintained with a constant flow of water of controlled temperature. The difference in oxygen tension between inlet and outlet water was monitored continuously by an oxygen electrode; sampling was mainly of the outlet water with occasional sampling of inlet water by means of a timing device and sampling valve (Fig. 1 a).

By reducing the water flow to zero and closing the respirometer box, determinations of the rate of fall in oxygen tension enabled to be calculated (Fig. 1b).

In this method fish remained in the same respirometer box but with the lid removed so that gaseous exchange could take place at the air-water interface. The rate of decline in recorded by an electrode was lessened because of diffusion of oxygen into the water from the air. Calculations taking this into account make it possible to estimate the oxygen consumption of the fish (Tamura & Morooka, 1973). Torpedo was very suitable for this method for, although it remained stationary, the inspiratory current at the spiracles kept the water mixed (Fig. 1 c).

Following anaesthesia with MS 222, cannulation was carried out in order to record pressure changes in the oro-branchial and parabranchial cavities. The cannulae were brought out on the dorsal surface of the fish. Sampling of water was also done using these cannulae.

Blood samples were obtained following cannulation of a branch of the anterior mesenteric artery. The abdominal cavity was opened by a median incision and the viscera displaced to one side. A heparinized cannula was inserted into a small branch and passed into the dorsal aorta. The cannula was brought out inside a thicker tubing (pp. 120) inserted through the dorsal musculature. The abdominal incision was carefully stitched so as to prevent the entry of water. The fish survived these operations very well and there were very few losses for at least 1 week following the operation. Blood samples could be taken without disturbing the fish in any way. Blood pH, P_o and other parameters were measured on samples taken during hypoxia and other experiments using a Radiometer BMS 2 and oxygen electrode assembly. Oxygen dissociation curves were determined using a mixing method (Hughes, Palacios & Palomeque, 1975).

Pressure changes were recorded using Sanborn 268 differential manometers and Beckman LVDT couplers. They were displayed on a Tektronix 564 oscilloscope and a Beckman 2-channel recorder. Simultaneous tape-recordings of 4-channels enabled fuller analysis and recordings to be made if required at a later date. Analysis of the cardio/ventilatory coupling was carried out after recording the data on punched-tape using a Hewlett Packard 9810A calculator (see Hughes & Adeney, 1977).

As in Raia the main inspiratory water current enters via the dorsal spiracles, which usually open synchronously during each ventilation cycle. The external openings are much smaller than in many other rays and in some specimens there are periods when only a single spiracle is functional ; such asymmetry was also observed in Raia clavata(Hughes, 1960). The time-course of pressure changes in the oro-branchial and para-branchial cavities showed positive phases of relatively short duration. Expansion appears to be a largely passive process and occupies about four times the duration of the active contraction phase. Expansion of the cavities is associated with a reduction in pressure and the time course of changes in the parabranchial cavities shows a more prolonged negativity (Fig. 2). The waveforms in the oro-branchial cavity slightly precede those in the parabranchial cavities. As in other fishes the movement of water across the gills is due to positive pressure pump and negative suction pump phases. The direct recordings of the differential pressure show similar waveforms to those derived for Raia clavata; the maximum differential pressure being very small (0 ·1−0·2 cm H₂O). As in Raia, there does not seem to be a reversal phase. During hyperventilation the amplitude of the pressure waveforms increases markedly (Fig. 9) and the differential pressure amplitude may be as great as 0·4 cm H₂O.

Most experiments were carried out using the continuous flow method and results on different sizes of fish are plotted on log-log co-ordinates in Fig. 4. It is clear that the oxygen consumption is almost directly proportional to body mass. Each of the points plotted on this curve was derived from recordings such as those shown in Fig. 3(b). Following an initial equilibration period the oxygen consumption shows some minor fluctuation but is generally fairly steady (Fig. 3a). Mean values were taken at this steady lower level of oxygen uptake, occasional higher values being omitted from calculations of the mean. It is apparent that this resting level of oxygen consumption may be maintained over several days and is relatively low when compared with figures in the literature for other elasmobranchs (Table 1).

Comparisons of this method with the other two indicated that closed chamber respirometry may produce higher values, perhaps due to the accumulation of carbon dioxide and other products during the experiment. In general, however, the results compared quite well with those of the continuous flow method. The application of the third method to fish of this kind for the first time is of special interest and gave quite similar results. This method clearly has advantages over the closed respirometer in that there is less accumulation of carbon dioxide, but as in that method there will be some accumulation of other waste products which would not be present with the continuous flow method. Some figures obtained for other species using this method are given in Table 2.

Several studies of these rhythms in Torpedo were made many years ago (Schoenlein & Willem, 1894; Schoenlein, 1895; Willem, 1921), and it was therefore of special interest to apply the method (Hughes, 1972) developed for dogfish to this species. Tape recordings were made at regular intervals over a period of several days. The recordings were analysed as interval histograms (Fig. 10) and event correlograms plotted out on polar co-ordinates to show the phase angle and percentage coupling between the two rhythms. These plots indicate the position of the heart beat between successive oro-branchial waveforms (Figs. 5, 6). In some cases the heart beat tends to occur in certain phases but at other times the rhythms seem much more independent and the coupling is very low.

Under normoxic conditions the results were similar to those found with dogfish, i.e. individual specimens showed different degrees of coupling and phase angles which also varied with time. Therefore it is not possible to make one generalization regarding the particular relationship between the two rhythms which is applicable at all times and for all individuals. Analysis of the recordings suggested that greater coupling occurred at lower ventilatory frequencies (Fig. 7) and also that the coupling was greater when the ratio between two frequencies was simpler, i.e. towards whole number ratios (Fig-8).

The most obvious response of Torpedo to a lowering of oxygen tension of the inspired water was an increase in frequency of ventilation which was associated with a greater amplitude of these movements (Fig. 9). Analysis of interval histograms during such hypoxia indicated not only a reduction in the interval between successive respiratory cycles, but also a lessening in the range of these intervals (Fig. 10). As the rhythm became more rapid the cardiac rhythm was not so regular both during normoxia and hypoxia and showed relatively small changes in the mean interval. It is only in extremely prolonged hypoxia that a marked lengthening of the interval between the heart beats was found. Where such a bradycardia was not so well defined there was often greater irregularity in the cardiac rhythm (Fig. 11).

Analysis of the cardio-ventilatory coupling during hypoxia was carried out in detail and a comparison was made between the results of the automatic computer analysis and a more laborious cycle-by-cycle analysis of the same recordings. Results of the latter are shown in Fig. 12, where the position of each heart beat in a ventilator cycle is given. The overall effect of hypoxia in this particular fish was a significant reduction in coupling which increased again on restoration of normal levels of inspired oxygen. This type of result was found in about 40% of the fish whereas the expected increase in coupling was found in only about 20%. It was only in later experiments when Torpedo was subjected to prolonged periods at such low inspired oxygen tensions (less than 20 mmHg, for more than 4 h) that definite evidence of bradycardia and some associated increase in percentage coupling between the two rhythms was obtained. Subsequent studies have demonstrated the presence of adaptations at a biochemical level during such prolonged hypoxia (Hughes & Johnston, 1977).

Although relationships for various properties of the blood and of the water blood/ barrier were investigated, no definite relationship with body weight was established.

As in previous work on the thornback ray (Hughes & Wood, 1974), most of the oxygen dissociation curves were determined using whole blood samples and a mixing method for determining the P₂₀, P₅₀, P₇₀, P₈₀, etc. Because of changes which occur in the properties of blood following its removal from the fish, only two determinations were made on each of the blood samples, as this can be carried out within half an hour of sampling. The oxygen dissociation curves for whole blood, shown in Fig. 14, were obtained by this method using blood equilibrated at two different temperatures. The mean P₅₀ at normal physiological conditions (15 °C, pH 7·8) was 20-2 mmHg. The oxygen-carrying capacity is small (3−4 ml O₂/100 ml blood) and consequently a fair proportion of the oxygen carried in the blood is in solution and this has not been subtracted from the data plotted in Fig. 14. The Bohr factor Δ log P₆₀/Δ pH, although small, was quite definite, having a value of −0·32. The in vivo arterial blood oxygen tension under resting normoxic conditions was usually about 70 mmHg and a pH of 7·82 at 15°C. More detailed studies on the characteristics of blood will be reported elsewhere.

Results of several types of investigation on the respiration of Torpedo marmorata confirm the view that it is a species adapted to a sluggish mode of life. This is in general agreement with the limited information available about its life habits but there is evidence that other species of Torpedo(e.g. T. nobiliana) may be more active. The oxygen consumption of T. marmorata, measured by three different methods, is lower than that of other elasmobranchs of comparable size, as is the surface area of its gills. Nevertheless, the ratio between these two figures (i.e. the amount of oxygen transferred : unit area) is similar to that obtained for other species (Table 4).

The general rested behaviour of Torpedo proved convenient in experiments comparing three different methods for measuring oxygen consumption. The locomotory activity was almost zero in all cases and the levels of O₂ uptake remain constant for long periods. The continuous flow method is preferred but the other two methods are possible with less equipment. The direction of water flow, in at the spiracles and out of the gill slits, created a current which gave good mixing of the water. This was especially valuable in the use of Morooka’s open respirometer method. Perhaps this method could have wider application, especially in the field, where it might be used to measure the oxygen uptake of single specimens of rare species, e.g. Latimeria, simply by following changes in oxygen tension of the water in a tank where it was kept alive. This method has the advantage over closed respirometry that the oxygen level does not fall to the same extent. Temperature control is important when using oxygen electrodes, but can be difficult in field conditions. No attempt was made to investigate the effect of environmental changes on oxygen uptake. It would be interesting, for example, to know the effects of discharge of the electric organs upon resting oxygen consumption.

Ventilation is achieved by mechanisms essentially the same as those in Raia(Hughes, 1960) but in this case the spiracular openings are much smaller. The pressure changes have similar relationships but the oro-branchial pump appears to be relatively more important although the parabranchial cavities clearly perform an important function in sucking water through the gills.

The low frequency and amplitude of the ventilatory movements are associated with a small differential pressure across the gills. The latter present a fairly coarse sieve of relatively low resistance to water flow and consequently the ventilatory activity must be relatively economic in energy requirements. The regular and low frequency (10−15 min) of the cardiac rhythm probably has a similar advantage. These features of the respiratory system are similar to those that are often found in other bottomliving fishes. The respiratory properties of the blood reported here are also typical of fish having relatively sluggish habits. The oxygen-carrying capacity is small and the high affinity blood has its steepest part at oxygen tensions below 20 mmHg. Under normoxic conditions the arterial blood is more than 90 % saturated although the is only about 70 mmHg.

During hypoxia the initial response of Torpedo shows a definite increase in both frequency and amplitude of the ventilatory movements. Changes in frequency are often quite marked (> 100%), which contrasts with its relative constancy in other elasmobranchs such as the dogfish (Ogden, 1945; Satchell, 1961 ; Hughes & Umezawa, 1968a; Piiper, Baumgarten & Meyer, 1970). Unlike other elasmobranchs (Satchell, 1961 ; Piiper et al. 1970; Hughes, 1973) bradycardia was initially thought to be absent in Torpedo but of course it is more difficult to recognise changes at such low frequencies. Under more extreme hypoxia, both in the level of reduction and its duration, bradycardia becomes clearly recognizable below 20 mmHg. Under these more extreme conditions the ventilatory movements become reduced in frequency and amplitude and there follows a stage of intermittent ventilation before it ceases altogether. During the stage of periodic movements it is often observed that the fish produces a ‘spout’ of water through the spiracles which is succeeded by several ventilatory movements of low amplitude.

These observations showing that the ventilatory rhythm is not so constant support the view of Willem (1921) who disagreed with the statement of his former collaborator (Schoenlein, 1895) that the two rhythms moved together so that the pulse and breathing frequency always changed in the same direction. However, the relationship between the two rhythms is by no means always a simple ratio, whether this be in the form of a synchronism (i.e. 1:1) or the modified form of synchronism (2:1, 3:1, 4:1) suggested by Willem (1921). However, the coupling tends to be highest when the two rhythms approach such simple ratios. Under normoxic conditions some specimens show a relatively high coupling, but under hypoxia when the ventilatory movements are accelerated and coupling tends to fall there is no corresponding change in the heart beat frequency. Coupling decreases during hypoxia more often than it increases, a response opposite to that found in rainbow trout (Hughes, 1973). Under the more prolonged and deeper hypoxia there is some indication of an increase in coupling in some specimens. It is of interest that Couvreur (1902) was unable to find synchronism between movements of the heart and respiration similar to those reported for teleosts and cyclostomes.

The interpretation of the significance of relationships between heart and ventilatory rhythms is more difficult when the two rhythms have very different frequencies. An increased effectiveness of gas transfer would be expected when the flow rates of water and blood coincide. Such a relationship cannot possibly hold where the two frequencies differ very widely. If the heart has a lower frequency than the ventilation then it would clearly be advantageous for the periods of maximum blood flow to coincide with periods of maximum renewal of water at the gas exchange surface. The ventilatory cycles which do not coincide with the heart rhythm would therefore be expected to have a lower effectiveness. The analyses so far carried out are at a relatively crude level concerned with the bulk flows of the two media whereas what is really important from a functional point of view is the relationship between the flows of water and blood at the level of single secondary lamellae.

Preliminary bio-chemical analysis indicated an accumulation of lactate, during hypoxia and more detailed studies during prolonged experiments (Hughes & Johnston, 1977, 1978) have shown the presence of alternative metabolic pathways in addition to the normal pathways of glycolysis. A significant increase of blood succinate may be the result of amino-acid fermentation and has been recognized as an adaptation of a number of diving vertebrates (Hochachka et al. 1975). Anaerobic metabolism has been shown to occur in a number of freshwater fishes (Blazka, 1958; Johnston, 1975) and fish are known to live in environments where the oxygen content of the water is almost zero (Coulter, 1967). During the course of the hypoxia experiments it is unfortunate that no measurements were made of oxygen consumption, especially during recovery from the long hypoxia. From measurements during some more extended closed and open respirometer experiments there were indications that the oxygen uptake falls when the oxygen in the water had been reduced below 50 mmHg. In view of the adaptations at a biochemical level it is important to know whether this fish pays off its ‘oxygen debt’ by oxidizing succinate and/or lactate.

It would seem that Torpedo has a very high safety factor such that the oxygen tension in the water can be lowered a great deal before any emergency mechanisms are called into action. The initial stages of regulation would maintain respiratory homeostasis but once the capability of this mechanism has been surpassed the fish seems to accept a more passive role until, at even lower oxygen tensions, biochemical mechanisms come into action. An understanding of the overall adaptation of Torpedo must take into account all of these mechanisms which have so successfully become integrated that the fish can remain apparently passive in the face of such extreme environmental stresses which in most other fish would certainly initiate escape mechanisms.

T. marmorata generally migrates into the Bassin d’Arcachon during late summer when the water temperatures are fairly high. It is often found at low tide in shallow mud pools near to the oyster beds. In these pools, as well as the bottom of the Bassin, Torpedo is largely covered by mud with only the spiracles visible. If the electric organs are used to paralyse its prey then clearly it is a fish which is best able to survive by remaining stationary and quiet. Thus it seems very likely that at least some specimens would find themselves in environments in which the oxygen level was extremely low. Measurements would need to be made close to the spiracular openings if valid figures were to be obtained. Under other circumstances it seems probable that certain species of Torpedo can be more active, presumably migrating between different parts of the sea bed.

As always I am pleased to thank M. Boisseau and M. Cazaux and their staff at Arcachon for generous and friendly help. I was also fortunate to enjoy the enthusiastic research assistance of Richard Adeney and Julian O’Neill during most of this work. Funds were provided by NERC and the Browne Fund of the Royal Society.