ABSTRACT
The pumping mechanism maintaining water flow across the gills has been investigated in the trout using pressure transducers and ciné analysis (Hughes & Shelton, 1957, 1958). It was shown that the mechanism could be considered as being made up of a buccal pressure pump and opercular suction pumps, the two being separated functionally by the gill resistance. In this species, Salmo gairdneri, the relative contribution of the two pumps appeared more or less equal, as judged by the areas of the corresponding phases of the differential pressure across the gills. Further studies of the respiratory pumps with special reference to their muscular basis were made using electromyography (Ballintijn & Hughes, 1965) and emphasized the many couplings between the two pumps. Although several studies have been made on the influence of environmental changes on gas exchange (Randall, Holeton & Stevens, 1967), no detailed account has been published on the modifications that occur in gill ventilation mechanisms and interactions with cardiac cycles when trout are subjected to stress. The present series of experiments forms part of such a study in which electrocardiograms were obtained simultaneously with electrical recordings from parts of the respiratory neurones in the medulla, from some respiratory muscles, and with pressure recordings from the buccal and opercular cavities. All were carried out during subjection of rainbow trout to warming thermal stress. A preliminary report of this work has already been given (Roberts & Hughes, 1967).
MATERIALS AND METHODS
The rainbow trout, Salmo gairdneri, used in these experiments were obtained regularly from a hatchery at Nailsworth, Gloucestershire. Specimens were usually of 200−300 g. and were allowed to acclimate in holding tanks within the laboratory for at least 7 days at a temperature of 15°+ 1° C., but in some instances the acclimation temperature was 18° ± 1° C. (Fish I).
Each fish was anaesthetized in MS 222 (0·1 g./l.) before it was fixed in the experimental tank and was allowed to partially recover to a lighter level of anaesthesia after the electrodes and/or pressure needles had been placed in position. The fish was then secured lightly in a body clamp of a sling type (velcro plush and hooklet band in. wide), but more rigidly in a second clamp attached across the supra-orbital ridges of the frontal bones (Fig. 1B). The water in the experimental tank formed part of a closed circuit 5 1. in volume (Fig. 1 A). Circulation was maintained by an air-lift device at a rate sufficient to allow accurate temperature control as well as systematic variation of the temperature (1°C. change/3 min.). The air-lift circulation also served to maintain the dissolved oxygen content of the bath water at saturation regardless of temperature (O2 content determined by Winkler). Pressure changes were recorded through hypodermic needles (O.D. = 0·83 mm.) with holes in the side (Fig. 1), which were fitted on to lengths of lead tubing attached to Sanborn 268 B differential pressure transducers. In most experiments an arrangement (Fig. 1C) was made for the simultaneous recording of either the buccal and opercular pressures or the buccal and differential pressures.
In a typical experiment the fish was allowed to settle in the experimental tank for about 1 hr. at its acclimation temperature and under an anaesthetic level which kept the fish generally quiet but which did not block gentle sculling movements of the pectoral fins (50−60 mg. MS 222/I.). Recordings were made of the buccal, opercular and differential pressures, electrocardiograms and any other parameters used for that particular experiment. The water temperature was then gradually increased and recordings were made at appropriate times. Temperatures were not allowed to rise very much above a bath temperature of 29°C. before being returned to the original level at the same rate of cooling as for heating. In only a few experiments, however, was complete recovery possible once the brain temperature had been above 27° C. The warming rate of 1° C./3 min. was chosen to conform with warming rates frequently used in the determinations of lethal thermal limits (Fry, 1967) and to minimize possible thermal shock effects (faster warming) and significant thermal adaptations (slower warming). As a result, the deep body temperature was found to lag behind the increase in bath temperature during warming. Body temperatures were taken by means of a bead thermistor implanted in the cranial cavity so that the temperature recorded at any one moment most nearly matched that of the respiratory neurones in the medulla. Figure 2 was constructed in order to permit intercon-versions of temperatures at any time during a warming sequence and is based on parallel recording of water bath and cranial temperatures. Except where otherwise noted, all temperatures given in this paper are those of the water bath.
RESULTS
A. The typical differential pressure record
In the original work on trout (Hughes & Shelton 1957, 1958), only a single condenser manometer was available and it was necessary to superimpose successive recordings of the buccal and opercular curves in order to obtain the differential pressure from them. With the development of the liquid differential transducer it has become possible to record the differential pressure directly. Hence a series of experiments was carried out for this purpose and, as well, to compare such recordings with those obtained using the cannulation technique of Saunders (1961). It was found that differential curves of the type derived by Hughes and Shelton were the most common (Fig. 4 A) with both methods of recording but details such as the relative balance of the buccal and opercular phases varied more as between specimens than as between methods. Unlike the results obtained in earlier experiments, subsequent recordings using rainbow trout from the same hatchery have shown periodic ‘coughs’ especially during early stages of their acclimation to an experimental situation. Recordings of such coughs were obtained during electromyographic studies and more recent recordings have been obtained in experiments on suspended solid pollutants (Hughes, 1970).
Typical recordings are shown in Fig. 3. The buccal pressure pump and opercular suction pump phases can easily be distinguished and are separated by two transitional phases, one of which shows a clearly defined reversal. Similar recordings with the animal restrained in the clamps were obtained using either the cannulation tech nique or via the pressure needles. The same was true of the pressure changes recorded during the coughs so that any differences which appear in experiments with restrained fish and fish relatively unrestrained in a closed circulation bath cannot be ascribed to the recording method.
For the more detailed analysis of the pressure waveforms recorded in this way, the following methods were used: (i) the total amplitude of the opercular or buccal pressure was measured directly from the recordings; (ii) the differential record was divided into three parts as shown in Fig. 4B; during two of these (a, b) the buccal pressure exceeds that in the opercular cavity and the third (c) is formed by the reversal pressure. Under certain conditions the second transitional phase also shows a reversal and this will be referred to as a double reversal. From tracings of the records measurements were made of the area of these different phases, so as to give some indication of the relative contribution made by different parts of the respiratory cycle to the pressure gradient operating across the gill resistance to produce the water flow. A relationship between the differential pressure and ventilation volume has been demonstrated (Hughes & Shelton, 1958; Hughes & Saunders, 1970). The algebraic sum of the area beneath the differential curve multiplied by the frequency/ min. gives a differential pressure equivalent for the minute volume and is therefore a useful measure of the performance of the pumping mechanism during varying conditions, especially when these produce frequency change. The mean differential pressure is obtained when this equivalent [(a + b − c) × frequency] is divided by time expressed in the same units as for the differential area measurements.
B. The effect of temperature stress
The frequency of respiration
In all preparations it was found that an increase in temperature led to an increased frequency of pumping. There was some variation in the details of this relationship but a number of preparations suggested that the rate of rise in frequency steepened as the temperature reached about 24°C. This is the equivalent of a brain temperature (deep body or core temperature) close to 22° C. at the warming rate used.
Table 1 shows the average frequency of respiration at 16° and 26°C. for a number of preparations as calculated from the pressure recordings. The data of Fig. 5 is from an experimental series in which electrical activity of respiratory neurones in the medulla was recorded. Both sets of data refer to changes in frequency which occurred as the Water temperature increased. They show that there is a sizeable individual variation not only in the base frequency level at the commencement of the experiment but also in the percentage increase. This suggests, as do other characteristics of this and other data, that there may be a variety of ways in which a fish compensates for stress, e.g. changes in stroke volume can occur and alter minute volume even at a relatively constant frequency. The fact that the incremental change in ventilation rates as shown in Fig. 5 was found to be very small, between 15°C. and 21°C. (Q10 = 1·18) is one type of evidence. Another piece of evidence was the observed sharp rate change above 21° C. which may represent the upper limit to such other means of compensation and/or reflect the onset of reflex cardiac inhibition that becomes noticeable above 21° C. Value of Q10 at the higher ranges were 2·34 between 21°and 26° C. and 4·02 between 26° and 28° C. The calculated overall Q10 values found for the experiment summarized in Fig. 5 and Table 1, were only 1·72 and 1·81 respectively. Clearly several interacting phenomena are involved.
In those experiments in which the readings were continued as the temperature was reduced to normal levels it was found that the respiratory frequency was higher for a given temperature than during temperature increase (Fig. 6). This was partly due to the fact that with the warming and cooling sequence used (1°C./3 min.) thermal gradients develop between bath temperatures and brain temperatures of the fish. Another factor involved in the hysteresis might be due to the paying-off of an oxygen debt incurred during the temperature rise.
Effects on cardiac frequency
The same slow temperature rise between 15° and 21° C resulted in cardiac rate changes with a Q10 of 2·07. Between 21° and 26° C. the rate of change in cardiac frequency declined (Q10 – 1·49). Very marked deceleration in rates (negative temperature coefficients) occurred above 26° C. up to the maximum permitted (29°–30° C.). It was also at this ‘break point’ in temperature that the sharp rise in ventilation frequency was found to occur with further warming (Fig. 5).
Possibly this serves as a means for adjustment to the parallel decline in cardiac frequency. The ratios between cardiac and opercular frequencies determined from the data of Fig. 5 changed from 1:1·7 at 15°C. to 1: 1·2 or nearly equal rates at 21° C. The ratio returned again to 1:1·7 at 27°C. and finally reached 1:3 as the high-temperature bradycardia developed. Whether or not increasing stress tends to increase the degree of cardio-ventilatory coupling was not always obvious for reasons of experimental design and methods of recording. For example, the temperature rise used was continuous at 1° C./3 min., a procedure which precluded the accumulation of sufficient data at any given temperature. A tendency to phase-lock or synchronize over the temperature range when the two activities were similar (22° and 26·5° C.) was not evident in the data used for the preparations of Fig. 5, but in other cases there were indications that it occurred.
Abnormalities in the electrocardiogram, such as arrhythmias, prolongation in P-R intervals and dissociations of atrial and ventricular beats, were not found to occur until very high brain temperatures had been reached (above 25° C.) and progressive failure in ventilation became obvious. The latter was usually identified by a rapid rise in the incidence of atypical respiratory movements of a spasmodic nature at about 26° C. Generally the cardiac abnormalities were not seen until regular ventilation cycles had ceased and when hypoxia was likely to be severe.
Effect on the buccal and opercular pressure
An increase in temperature was accompanied by significant increases in the amplitude of both the buccal and opercular pressure waveforms (Fig. 6) which tended to be more marked for the opercular pressure in most preparations. At the higher temperatures (above 23°C.) the pressure amplitudes fall, that of the buccal pressure begins at a slightly lower temperature than the opercular pressure. During recovery the opercular pressure does not have such a great amplitude at a given temperature as during the warming period (Fig. 6 a, b). Such differences in amplitude between the increasing and decreasing phase of temperature changes are not usually so clear for the buccal pressure. During recovery the buccal curve may or may not be greater in amplitude than during stress and seems to change relatively more than the opercular pressure during this period. In some cases the buccal pressure increased in amplitude to a far greater extent during recovery than did the opercular pressure (Fig. 6I).
Apart from increases in amplitude the increase in water temperature may also give rise to changes in form. At intermediate temperatures the pressure waveforms generally become more regular and constant in form, but above temperatures of 21−22°C., the buccal waveform sometimes shows an increase in its negative phase which may be associated with a fall in gill resistance and consequently a greater effect of opercular suction on the pressure within the buccal cavity. There are also frequent abnormalities, such as superimposed biphasic waveforms, which suggest alterations in the coupling between the two pumps, especially as they are accompanied by the appearance of double reversals in the differential curves.
It is difficult to make overall generalizations regarding pressure waveforms, for not only do they vary under normal conditions, but, because of the relatively balanced role of the two pumps in the trout, there are several apparently equally effective ways in which they can respond to the temperature stress. The examples plotted in Fig. 6 I and II illustrate this variability.
Effect on the differential pressure curves
There tend to be effects on the opercular suction and buccal pressure pump phases of the differential curve which are related to changes in amplitude of the individual waveforms. The transitional phases frequently became more marked, and this was especially true of the reversal phase in nearly all preparations. As the temperature was raised, there was an increase in the proportion of the respiratory cycle during which the respiratory pressure showed a reversal. This varied; for example, the percentage reversal for fish G was 7·1% at 15° C., rising to 10% at 22·5° C. and 28% at 25·5° C. At the higher water temperatures there was often a double reversal which contributed to the greater total reversal time. In fish I, however, which was generally a good preparation with well balanced buccal and opercular pressures, at 15·6° C. the reversal time was 12·5% but it had fallen to 9·7% at 23·5°C. However, at 26·4° C. there was a 25% reversal time with double reversals. On cooling the same fish, double reversals fell out when the temperature had fallen to 23° C. This fish survived a water temperature of 22·9°C. In other fish the reversal time at normal temperatures (15°C.) might be as high as 29%. In this one case (fish L) the reversal time with warming rose to 39% at 21·5°C., but then dropped until the stress temperature approached lethal (about 29° C.) and ‘coughing’ began. There were instances during the warming stress when for short periods the pumping efficiency appeared to increase, as in fish L at 25−3°C. when the reversal time dropped to 5·7% and no longer showed double reversals. It is worthy of note that the buccal pump was dominant in this fish throughout ; such inequalities between the pumps seem to increase the occurrence of double reversals.
It was found that there was usually an increase in the area beneath the differential pressure curve considered in relation to a single cycle or on an overall basis. The differential area/stroke (a+b—c) usually increased with temperature and then declined as it reached higher levels (about 21° C.) as illustrated for fishes I and J in Fig. 7.
There was also a tendency for the total differential area/unit time or the mean differential pressure (Fig. 6C) to increase as the frequency of respiration in creased, but at higher temperatures (and frequencies) there was once more a decline coupled with the increased double reversal phases mentioned above.
Under conditions of greatest temperature stress the fish makes very rapid pumping movements of small amplitude and probably pumps a relatively small total volume of water across the gills. At intermediate temperature stresses there were indications that the fish was compensating by increased levels and efficiency of pumping. One sign of this was taken to be the relative stability of the mean pressure below temperatures considered to be severely stressful, i.e. about 26° C. (as shown, for example, by fish I, etc.) (Figs. 6 and 8). Also the variation found for the patterns of change in the pressure curves with warming was considerable among the seven trout used in this part of the study. This suggests that differing combinations of frequency and amplitude are utilized by individual fish in the operation of the ventilatory pumps. Although it is obvious that in any one fish, buccal and opercular pumps cannot vary separately in frequency, pressures developed during ventilation can and do differ greatly as between the two cavities. This means that for any one fish the energy needed to pass water through the gills can be expended equally by both pumps or predominantly by one of them. Figure 7 illustrates two cases. In trout I the opercular pump dominated until pumping efficiency fell to very low levels at about 23 ° C. (CT). In trout J the buccal pump dominated for the entire warming period. These results confirm those obtained from measurements of the amplitude of the individual pressure waveforms. The declines in differential pressures shown by these two fish with severe thermal stress were also found with the other animals in the experimental series (Fig. 8). Because the pressure changes have such low amplitudes at higher temperatures (e.g. Fig. 3B and 4B), and may also change in form, it was difficult to separate the two phases generally ascribed to the buccal and opercular pumps. The independence of the two components of the differential curve may be greater in the trout than in other fish because of the lesser degree of coupling between the two pumps as has been indicated in the recent comparison made by Ballintijn (1969) of the carp and trout.
DISCUSSION
When a fish is subjected to thermal stress as in the present series of experiments, the response of the respiratory and cardiovascular systems is complex because of the varied effect of the environmental change. There is an overall effect on the metabolic level, as with terrestrial animals, and such studies have been carried out for brook trout by Job (1955), and Beamish (1964). It is noticeable (Fig. 9) that the Q10 of oxygen uptake declines at higher temperatures, an effect similar to that which was noted by Rao & Bullock (1954) for a wide variety of fish and other poikilotherms. In air-breathing animals the direct effect of temperature changes on the respiratory environment is not very significant, but for an aquatic animal there is a very marked effect on important physical characteristics of the water. These include its viscosity and density, and from the point of view of gas exchange, the solubility and diffusion rate of oxygen in water. Changes in these parameters with temperature increases between 5° and 25° C. are summarized in Table 2 and Fig. 9.
The lowering of viscosity must certainly reduce gill resistance and ventilatory work and thus partially compensate for the increased ventilation necessitated by the decrease in oxygen content resulting from a fall in solubility (Hughes, 1963). Yet, because the percentage utilization typically falls with increased flow (Saunders, 1961 ; Hughes & Shelton, 1962), this advantage may be effectively masked. At first sight, the rise in oxygen diffusion rate with temperature (3 %/° C.), might be expected to aid gas exchange, but this will not be so great because gas transfer involves a permeation coefficient and that of Krogh is most often applied in such circumstances (Steen & Kruysse, 1964; Hughes, 1966). From Table 2 and Fig. 9 it can be seen that the product of a and D, which is usually taken as equivalent to Krogh’s permeation coefficient (D′) (Radford, 1964), increases relatively slightly over this temperature range. Consequently facilitaition of gas exchange with a rise in temperature is not so great as expected. Thus the effect of temperature on the solubility of oxygen influences oxygen uptake adversely in two ways: first, by reducing the effective diffusion coefficient and secondly by reducing the total oxygen content of the water and so necessitating greater ventilation. Increasing ventilation will, however, reduce the thickness of the water film over the secondary lamella surface, which provides the major resistance to gas exchange (Hills & Hughes, 1970). It is difficult to draw up a complete balance sheet for these different factors but clearly the increased metabolic demand presents a considerable load for the respiratory pumps.
The results summarized in this paper show how the fish responds by an increase not only in the frequency of its ventilation and heart movements but also in the amplitude of pressure changes in the respiratory cavities which must result in an increased stroke volume. Frequency is a relatively easy parameter to measure and shows significant effects of temperature on both the ventilatory and cardiac pumps. The observed effects upon heart rate were as expected (Q10 of about 2 until 21° C.) but rate changes in ventilation proved not to be so predictable. Q10 values for the change in ventilation frequency ranged from nearly 1 between 15° and 21° C., to a high value of 4·02 between 26° and 28° C. It is of interest that changes in stroke volume seem more important than frequency increases during the initial heat stress because during hypoxia frequency also increases, mainly during the more extreme stresses (Hughes & Saunders, 1970). One possible interpretation of these effects is that the initial rises in ventilation and of cardiac frequency are sufficient to provide enough oxygen to meet the raised metabolic demands, possibly by adjustments in the flow patterns of water and blood across the gills. But added stress leads to ventilation becoming inadequate and possibly the resulting depletion of oxygen in the blood evokes the bradycardia. This cardiac ‘braking’ is paralleled by a sharp rise in ventilation frequency, perhaps as a compensatory consequence. The inadequacies of the ventila-tory pumps can be partly offset, as indicated by a rise in brain levels, by a stream of water into the mouth (Roberts & Hughes, 1967).
The precise mechanisms mediating bradycardia are only partly understood. It seems clear that in trout and some other teleosts the efferent pathway for both behaviourally and physiologically induced bradycardia is via vagal cholinergic fibres with endings on the heart (Labat, 1966; Randall, 1968; Roberts, 1968). In some species the bradycardia which occurs at high temperatures is partially relieved by bilateral vagotomy (Labat, 1966) or by pre-treating the fish with pericardial injections of atropine (Roberts, 1968). It may be assumed that the cardio-inhibition found above 26° C. in these experiments is the final result of hypoxia related to failure of the cardioventilatory system to maintain adequate oxygen levels at the appropriate receptors (see Randall & Smith, 1967). If this assumption is correct in the context of the present work, receptor detection of oxygen must be occurring centrally (e.g. in the brain or cardiovascular system) rather than peripherally in the branchial system, for air saturation was maintained in the water passing over the gills at all temperatures. Yet results of Randall & Smith (1967), also with rainbow trout, contrast in that peripheral detection probably induces the hypoxic reflex bradycardia that occurs in water with less than 80−100 mm. Evidence for the functioning of both central and peripheral reception of changes controlling ventilation has been found in Callionymus (Hughes & Ballintijn, 1968).
Still another possibility exists which seems not to have been tested, and that is that the bradycardia results from the input of thermal receptors which can modulate the vagomotor centres. The probability is that many afferent pathways exist, not all related to oxygen detection, which can induce reflex bradycardia via vagal action. This becomes more evident when consideration is given to the observation that immediate bradycardia results from lifting fish out of water (Serfaty & Raynaud, 1957), and as a result of visual disturbance (Labat, 1966; Roberts, 1968).
SUMMARY
Trout subjected to changes in water temperature (1°/3 min.) between 15° and 30° C. showed a number of responses in their ventilation and cardiac mechanisms.
Ventilation rate increased slowly over the range 15° − 21°C., but increased more rapidly at higher temperatures (21° − 26°, Q10 = 2 · 34; 26° − 28°, Q10 = 4 · 02). Cardiac frequency fell markedly about 26°C., and this bradycardia suggests that above this temperature the ventilation mechanism is inadequate to maintain a sufficient level of blood Po. Possibly this insufficiency results from a failure of the pumping mechanism to increase or even maintain a large minute volume at high frequencies.
Pressure recordings indicate those parts of the ventilatory mechanism which are mainly involved in these responses. Increases in the buccal, opercular, and mean differential pressures indicated that the volume of water pumped across the gills increased during the initial stages of warming and only at higher temperatures did frequency become involved.
Variability in the balance between the buccal and opercular pumps among individual trout becomes even more apparent under temperature stress, as in some cases the opercular pumps seem to be mainly involved in the increased ventilation but not always.
The action of the buccal and opercular pumps seems well co-ordinated, especially at intermediate temperatures, but serious uncouplings occur at higher temperatures as indicated by the shape of the pressure waveforms and especially the appearance of double reversals in the differential pressure curve.
The relationship between cardiac and ventilatory cycles was not studied in great detail, but there were certainly indications of changes in coupling. However, it was clear that the two rhythms do not always become phase-locked even at high temperatures.
The effect of changing temperature on the physical environment of the fish is discussed in relation to these observed responses.
ACKNOWLEDGEMENTS
Equipment was provided from a Natural Environment Research Council grant to G.M.H. J.L.R. was supported by Special Fellowship GM2531 of the U.S. Public Health Service.