Effects of Temperature on Branchial Sodium-Exchange and Extrusion Mechanisms in the Seawater-Adapted Flounder Platichthys Flesus L

Changes of internal levels of sodium and chloride may reflect changes of either electrolyte or water balance. Effects of temperature on water balance are described in the studies by Mackay & Beatty (1968) on kidney function of the white sucker fish and by Evans (1969) on the changes in the diffusional water permeability of the gills of various teleosts. Very recently, R. Motais & J. Isaia (in preparation) have compared the effects of temperature on the exchange of sodium and water across the gills of various teleosts, especially the European eel.

The present study concerns temperature effects on branchial sodium transfer in the seawater-adapted flounder, Platichthys flesus. Isotopic techniques have made possible a reassessment of the ionic transfer mechanisms in the gills (see reviews by Maetz, 1964; Motais, 1967; Potts, 1968; Maetz, 1971). The flounder is one of the most investigated fish in this respect. Since the drinking rate is only some 200 μlh⁻¹ (100g)⁻¹, net absorption of sodium by the gut which is compensated by the extrusion mechanism in the gills is at most 100 μ-equiv h ⁻¹ (100 g)⁻¹ (Motais, Isaia, Rankin & Maetz, 1969; Maetz, 1969). Net sodium excretion by the gill corresponds to sodium efllux being higher than sodium influx. The unidirectional fluxes across the gills are large, however, as shown by the rapid sodium turnover rate measured isotopically, up to 50% of the internal sodium being renewed per hour. Thus the branchial sodium efflux is 2600 μ-equiv as against 2500 μ-equiv h⁻¹ (100 g) ⁻¹ for influx.

Rapid transfer experiments of seawater-adapted flounder to media of various salinities and to fresh water showed that the sodium influx and 85% of the efflux are dependent upon the external sodium concentration. Motais, Garcia Romeu & Maetz (1966) suggested that this part of the exchange flux is due to exchange-diffusion. The remaining 15% corresponds to passive diffusional flows and active extrusion. According to Ussing (1960), exchange-diffusion involves a sodium carrier, shuttling sodium to and fro across the membrane, irrespective of the electro-chemical gradient, The energy requirement of such a mechanism is theoretically small and it should be relatively independent of temperature. The initial purpose of the work presented here was to assess the temperature dependence of the exchange-diffusion component of the sodium transfer mechanism in the gills.

More recently, our understanding of gill mechanisms has progressed in two directions. First, it was observed that rapid-transfer experiments involve also electrical potential changes across the gill and notably potential reversal upon transfer from sea water to fresh water (Maetz, 1971). It is therefore probable that sodium exchange by way of passive flows along the electro-chemical gradient is a more important component than was hitherto suspected. The temperature dependence of this component is also expected to be small. Secondly, rapid-transfer experiments of flounder into potassium chloride solutions showed that up to 50% of the sodium efflux is dependent on external potassium concentration (Maetz, 1969). It was suggested that the branchial sodium-extrusion mechanism is effected by a Na/K exchange across the gill, presumably via Na-K-dependent ATPase. The present paper demonstrates that this Na/K exchange is much more temperature-dependent than the remaining Na/Na exchange, whether this takes place by simple or by exchange diffusion.

Flounders were collected in autumn at the mouth of the River Loire and brought by air to the laboratory. They were kept in running sea water (sodium concentration 520-570 μ-equiv/l) for at least a month before use. The temperature of the water was 16 ± 2 °C. The fish weighed between 75 and 325 g.

From this stock, some fish were adapted for from 1 week up to 4 months in closed-circuit seawater aquaria at 6 ± 2 °C. Survival was excellent. Other fish were kept in heated aquaria with running sea water at 23 °C (survival about 1 week) or 26 °C (survival a few days). The experiments were made between October and April over 3 years.

Several types of experiments were performed. Some fish were studied at the same temperature as that to which they were pre-adapted. The following measurements were made: sodium turnover rate, sodium efflux and its various components, drinking rate and muscle ionic content. The methodology allowed sodium space to be determined at the same time as the sodium turnover rate. In other experiments, the effects of abrupt temperature changes on sodium efflux and its various components were studied. Details of the various measurements made are as follows.

Internal sodium was tagged by intraperitoneal injection of ²⁴Na Cl (isotonic saline solution) in known quantity (about 200 μc per fish). The fish was then transferred immediately to a thermostated closed-circuit bath of known volume (usually 700 ml per fish) and the appearance of the tracer in the external medium was followed by a gamma-flow detector included in the closed circuit. A digital printout of the radioactivity flowing through the counter was obtained every 5 min (for details, see Motais, 1967; Tanguy, 1970). Samples of the external fluid were also taken at regular intervals and their gamma activities were compared to both the total amount of radioactivity injected and to the activity of a plasma sample taken at the end of the experiment. These comparisons were made using a well counter. The duration of the turnover experiment was generally 6 h.

Turnover rates were calculated from the rate constants of either the loss of ²⁴Na from the fish (Potts & Evans, 1967) or the rate of appearance of ²⁴Na in the external medium (Motais, 1967). Both the disappearance and appearance rate of isotope behaved as a simple exponential function, indicating that the system can be viewed as a simple two-compartment model. The equations were of the form Q_1nt = Q₀ e^−λt for disappearance and Q_0xt= Q_max (I–e^−λt) for appearance of isotope, where Q_1nt is the radioactivity remaining in the fish as a function of time, Q_o is the radioactivity injected into the fish at zero time, is the radioactivity which has appeared in the external bath as a function of time, while is the total radioactivity recovered in the bath after infinite time. All activities are expressed in counts per min (cpm). The half-times were determined graphically and the rate constant λ was obtained from ⁠, the turnover rates being given as % h ⁻¹ and being expressed in hours.

As shown in Table 1, both methods gave similar results. In Table 2 the means of the values given by the two methods are given.

The turnover experiments permitted simultaneous measurement of the internal sodium space. At the end of the experiment the radioactivity remaining in the fish (i.e. injected radioactivity minus radioactivity released into the external bath, in cpm per 100 g) was divided by the radioactivity of the plasma (in cpm ml ⁻¹) giving the sodium space in ml (100 g) ⁻¹.

While measurements of turnover rate and of internal sodium space were made in animals adapted at their environmental temperature, sodium efflux and its various components were evaluated not only at the adaptation temperature but also in experiments involving abrupt temperature changes.

Fishes were injected with ²⁴Na (approx. 200 pc per fish) and 30 min were allowed for distribution of the isotope throughout the internal sodium compartment. When it was required to test the effects of temperature on sodium efflux, the fishes were placed consecutively in two aquaria containing 700 ml sea water but at two different temperatures. The temperature difference was ± 5 or ±10 °C. The aquaria were in closed circuit with a pump, a temperature regulator and a gamma-flow counter. The appearance rates of ²⁴Na were recorded for 25 – 30 min first at the temperature of adaptation and then at higher or lower temperature and finally again at the original tempera-ture. At the end of the experiment blood samples were taken and the radioactivity of the plasma was measured (in a well counter) and compared to that of samples of the external bath taken at the end of each efflux period.

The effects of temperature on efflux were assessed by comparing the appearance rates of ²⁴Na taking into account the difference in efficiency of the flow-counters interposed in the circuits. Fig. 1 illustrates typical experiments of this kind, showing that the appearance is linear for the duration of the test. The slope observed during the second control period is generally comparable within ± 10% to that of the first control period, showing that the decrease of the internal specific activity is negligible and that the temperature effect is reversible. Therefore the effluxes obtained in the middle period were compared to the mean effluxes obtained in the two control periods.

The absolute value of the sodium efflux was calculated by the methods given by Maetz et al. (1967), which consists in dividing the appearance rate (in cpm h⁻¹100 g⁻¹) by the specific activity of internal sodium, using the values obtained by counting aquarium samples and plasma samples in the well counter, the plasma sodium level being measured by flame photometry. The internal specific activity given by the plasma sample taken at the end of the experiment was corrected by taking into account the radioactivity lost by the fish during each period. The sodium space is assumed to be 30 ml per 100 g for fish adapted to 16 °C and 50 ml (see Results) for fish adapted to 6 °C. The internal sodium level is assumed to be constant during the duration of the temperature test.

In some tests involving transfers to the higher temperature range, 23–26 °C, the duration of the successive periods was reduced to 10–12 min, since these high temperatures had deleterious effects on the gill mechanisms. In these experiments the radioactive efflux was recorded every 2 min to improve accuracy.

According to Motais et al. (1966) and Motais (1967), about 85% of the sodium efflux in the flounder is dependent on the external sodium concentration, presumably by way of a coupling between influx and efflux of sodium through a common exchangediffusion carrier. The degree of dependency of the sodium efflux upon the external sodium concentration was studied in fish adapted to 6 °C. Rapid transfers from sea water (SW) into fresh water (FW) and into various dilutions of SW with FW (1/2, 1/4, 1/8 SW) as well as into hypersaline SW, prepared by addition of NaCl to SW (final concentration: 855 m-equiv) were performed. The relative appearance rate of ²⁴Na is given by the comparison of the slopes obtained during the experimental period and the arithmetic mean of the slopes observed for the two control periods in SW as described for changes of sodium efflux with temperature.

As described by Maetz (1969), the ²⁴Na appearance rate of seawater-adapted flounders was compared in SW, FW and FW with 10 mm-KCl, the concentration of potassium found in SW. In some experiments the effects of 10 mm-NaCl were compared to those of KC1. This transfer technique is similar to that outlined above except that the efflux was monitored every minute to avoid decrease of the sodium efflux by secondary regulation (Motais, 1967), the fish remaining in the low-salinity media only for a total of 15 min. The experimental protocol was as follows. Sodium efflux was measured first in SW, then in FW after a 2 min rinse, then in FW with added potassium (FWK), and finally again in SW (SW₂). The order for the FW and FWK periods was alternated in successive experiments. A supplementary rinse of 1 min was made between FWK and FW to remove the potassium. In some experiments, especially when an additional test in FWNa was performed, the final SW₂ flux was smaller than the initial SW_X flux, presumably due to secondary regulation. The relative magnitude of the fluxes in FW, FWK and FWNa have been calculated assuming that adaptation proceeds linearly between SW₁ and SW₂.

The extra sodium efflux due to the addition of potassium to FW is assumed to be an index of the active sodium extrusion by the gill (Maetz, 1969). The extra sodium efflux due to the addition of 10 mm-NaCl is thought to represent the Na/Na exchange either by exchange diffusion or by passive diffusion along the electro-chemical gradient. The flux measured after transfer to FW is an index of the passive permeability of the gill to sodium. The purpose of these experiments was to compare the temperature-dependency of these various active or passive parameters.

Fish were tested at their adaptation temperatures (16 and 6 °C) or at a different temperature to study long-term and short-term effects of temperature changes on the various efflux parameters.

When a seawater-adapted fish is in a steady state, the amount of sodium swallowed by drinking and absorbed by the gut is equal to the amount of sodium extruded from the gills. Drinking rate gives therefore an indirect index of the sodium-extrusion activity. This rate was measured in flounders adapted at 6 °C for a week. The colloidal gold (¹⁹⁸Au) technique as described by Motais et al. (1969) was used. The fish were kept for 2 h in SW containing colloidal gold at I mc/l, after which they were transferred to ‘cold’ SW, anaesthetized with MS 222 at I g/l, and the gut was dissected out. The radioactivity accumulated in the gut was compared to that of an aliquot of the outside medium. Drinking rate is expressed in μl h⁻¹ (100 g)⁻¹.

It was found that internal sodium space increased as a result of adaptation to low temperature. An attempt was made to locate the extra sodium. Parietal muscle representing in flatfishes more than 50% of the body weight, appears to be the most likely site. Parietal muscle of fish kept at 6 °C for 4 months was examined for changes in cellular content of sodium and potassium.

The extracellular space of muscle was measured by injection of Na₂³⁵SO₄ into the peritoneal cavity of fish at least 8 h before sacrificing the animal. According to Rankin (1967) this delay is necessary to ensure homogeneous distribution of the extracellular label. A fraction of the muscle sample was homogenized in distilled water in a Teflon homogenizer after weighing, and the homogenate was left standing overnight to extract the cellular ionic content. A further fraction of the muscle sample was dried overnight at 95 °C to determine the water content. A blood sample was taken from the caudal artery just before killing the fish. The amounts of the ions Na, K and Cl in the plasma were measured either by flame photometry or by amperometric titration and compared to the amounts in the centrifuged muscle homogenate. The radiosulphate beta activity was measured in the plasma and in an aliquot of the homogenate, using a low-background Geiger counter. The radioactivity of the homogenate was assumed to be due to extracellular fluid, and ionic and radioactivity content of the plasma was taken as that of extracellular fluid. The electrolytes in the cells could therefore be calculated and expressed in μ-equiv per ml water content.

Table 2 summarizes the results. The plasma sodium level remains unchanged following adaptation to cold water, but disturbance of the sodium balance is nevertheless evident because of the 60 % increase of the sodium space (P < 0·02) after 7-12 days adaptation. The sodium space remains exceptionally high even after 2-4 days adaptation in cold water (P < 0·01). From these observations it may be concluded that the internal sodium mass (i.e. the product of the sodium space and plasma concentration) increases from about 5600-9100 μ-equiv per 100 g body weight. But as no further increase is observed after the initial period of adaptation, it must be assumed that the fish regains its mineral balance. Return to 16 °C is accompanied after 2 weeks by a 30% decrease of the sodium space (P < 0·05), showing that the perturbation is reversible.

The isotopic studies confirm further the disturbance of the mineral balance. In ‘cold’ flounders the sodium turnover rate is only about one third of that observed in ‘warm’ fish (P < 0·001), which indicates that sodium efflux is drastically reduced.

Moreover, adaptation of the fish in the cold for 2 to 4 months is not characterized by any compensatory change; but since the fish re-attains sodium balance, it is probable that the sodium influx is reduced to the same extent as sodium efflux. The diminution of the sodium exchange fluxes, as calculated from the product of the sodium turnover rate by the sodium mass, is by a factor 2. Direct measurement of the sodium efflux (see Table 2) confirms this value to be in the range of 1·7–2·2 (P < 0·001). Return to 16 °C is accompanied within 2 weeks by a prompt increase of the sodium turnover rate and sodium efflux (P < 0·001).

To what extent are the various components of the sodium efflux modified by the change of temperature?

Table 3 illustrates our attempts to compare the exchange-diffusion component of fishes adapted at 6 °C for 4 months to those studied by Motais et al. (1966) which were kept at 16 °C. The values of the sodium concentrations of the various transfer media are given in the table as well as the relative flux values with reference to sea water and the calculated absolute values of the efflux in the various media. In Fig. 2 the difference between the various fluxes and that found in FW (Δf_out) is given as a function of the external sodium concentrations of the transfer media. It may be seen that for this component at 16 °C as well as at 6 °C the curves suggest saturation kinetics and the presence of a carrier for sodium exchange. Fig. 3 shows a reciprocal plot of the same data indicating that the maximal flux increases by threefold, while the K_m remains unchanged. This might mean that the number of sites available for exchange-diffusion decreases with cold adaptation while their affinity remains unchanged.

In Table 4 it may be seen that the sodium-free efflux observed after transfer to FW (tr. FW) is reduced to the same extent as the total sodium efflux; that is, by a factor 2.

Indirect evidence suggests that the active extrusion mechanism is much more temperature-dependent than total sodium efflux or sodium-free efflux. Table 5 summarizes our data concerning the drinking rate measured in the fish adapted to 6 °C as compared to those given by Motais et al. (1969) for fish adapted to 16 °C. It may be seen that the drinking rate is reduced threefold against twofold for the sodium efflux or the passive sodium leak measured after transfer into FW. This ratio of three is probably a minimal estimate for the relative sodium extrusion activities of the gill in cold and warm environment since it is possible that in the cold the absorption efficiency of the gut is impaired and less sodium has to be excreted to compensate for sodium entry via the gut. Furthermore, the increase in sodium space suggests that the net extrusion rate is even slightly less than the rate of sodium absorption via the gut.

Our observations on the sodium efflux which is dependent upon external potassium confirm that the net extrusion mechanism is affected much more than the other components of sodium efflux. Table 4 collects our results concerning the fluxes observed upon transfer to FWK as compared to the fluxes after transfer to FW. When fish adapted at 16 °C are compared to those kept for 7 – 12 days at 6 °C, it may be seen that the FWK flux is reduced by a factor 6-8 and that the extra sodium efflux due to the addition of potassium (which is positive in fish adapted to 16 °C) is negative, although not different from zero in fish adapted to 6 °C. Figs. 4 and 5 illustrate this point (see the left-hand side of the figures concerning the transfer tests in the adapted fish). The potassium effect therefore has virtually vanished after adaptation in the cold environment.

Table 6 gives the results of our attempts to locate the extra sodium content of the fish in parietal muscle. It may be seen that adaptation to cold is accompanied by a great relative increase of the sodium content of the muscle. In absolute concentration this augmentation is, however, modest: 10 μequiv ml⁻¹. A concomitant decrease of the potassium content, also by 10 μequiv, is observed.

Here we describe the effects of abrupt temperature changes which presumably do not involve long-term compensatory mechanisms. From these experiments we hoped to gain insight into the biophysical mechanism of ion transfer, notably the activation energies involved respectively in the active and passive efflux components.

A series of rapid-transfer experiments were performed to test extensively the immediate effects of temperature change on total sodium efflux, which as seen before is mostly due to passive sodium transfers either by exchange-diffusion or by simple diffusion. Animals adapted to 16 °C for 2 months or more or to 6 °C for 7 – 12 days were used. Fig. 1 illustrates typical experiments and Table 7 shows all results. It may be seen that for the 16 – 6 – 16 °C temperature sequence the efflux is instantaneously and reversibly decreased to 63-5% of its control value. For the converse transfer (6–16– °C), an instantaneous and reversible increase to 159% of the control value is observed. In fact the two ratios are reciprocals of each other, showing that the immediate effects of temperature transfer are symmetrical. Table 7 also gives the results of a few 16-21-16 °C and x6-n-i6°C temperature-sequence experiments with similar results.

If the variation in sodium efflux found in abrupt transfer experiments is taken as the result of mere physico-chemical changes in the system transferring sodium across the gill, then these data can be used to calculate the activation energy of the process. In Fig. 6 the absolute values of the fluxes computed on the basis of a 2510 μ-equiv h⁻¹ (100 g)⁻¹ flux in fish adapted to 16 °C (see Table 2) and the flux ratios given in Table 7 are plotted using the well-known Arrhenius plot : log. flux against the reciprocal of the absolute temperature (in °K). According to the Arrhenius relation, we have

f₀ being the theoretical flux at o °K, R the gas constant (i.e. 1 · 987 g. cal mole⁻¹ (°K)⁻¹) and A the activation energy of the flux process expressed in kcal mole^-1. As may be seen in Fig. 6, the absolute fluxes observed for fish adapted to 16 °C transferred to 21, 11 and 6 °C fit a straight line suggesting that within these limits of temperature the Arrhenius law applies to the transfer process. From the slope of the line the activation energy A may be calculated: A = 7 kcal.mole⁻¹. In Fig. 6 the fluxes for the fish adapted to 6 °C transferred to 16 °C have also been plotted, calculated on the basis of a 1470 μ-equiv flux (Table 2) and the ratios of Table 7. The line thus obtained is parallel to the preceding one, as may be expected from the reciprocity of the processes. Thus the same activation energy is found.

Table 7 also summarizes our attempts to transfer fish at higher temperatures. Fig. 7 illustrates a typical experiment. It may be observed that in all experiments the return flux into 16 °C is smaller than the initial control flux. This irreversibility contrasts with the reversibility obtained in the lower temperature range. It suggests that some sort of damage to the transport system occurs at high temperature. When short and long exposures are compared, it is evident that the damage is less important after a short exposure to the hot environment. Absolute flux measurements made during these tests confirm the reduced efflux. For example, in the four experiments at 26 °C the return flux was 1974 ± 208 μ-equiv. h^{− 1} (100 g)^{− 1} as against 2750 ± 38 (P < o · o1) for the control period. The plasma concentration of these fish was 201 + 14 m-equiv. I^-1 at the end of this test, a value significantly higher than the control values for plasma sodium given in Table 2. This rapid shift in the internal sodium level strongly suggests impairment of the sodium-extrusion mechanism. Furthermore, the damage seems to have long-lasting effects, since one week later these fish kept at 16 °C showed a flux of 1812 ± 122 μ-equiv, significantly lower than controls (P < 0 · 02; see Table 2). A further experiment consisting in the transfer of these fish into 23 °C gave a final return flux of 492 ± 44 μ-equiv. This drastic reduction shows that further temperature shock has cumulative effects. Our observation that flounders do not survive transfer to 26 °C for more than a few days may well reflect osmoregulatory failure.

Table 4 summarizes our observations on fish adapted to 16 °C transferred to 6 °C and on fish adapted to 6 °C (for 7– 12 days) transferred to 16 °C. Higher-temperature transfers were not studied. These experiments were performed 2 years after those reported in Table 7. The most complete investigation was made on the fish adapted to 16 °C which were first tested to measure the components of the efflux at their own temperature of adaptation and then tested upon transfer to 6 °C. Fig. 4 illustrates a typical experiment. In this series of tests the total efflux is reduced by a factor 1·9 (A = 10· 6 kcal mole⁻¹). The residual flux after transfer to fresh water is reduced by a factor 1·3 which corresponds to a very low activation energy (4·5). The extra sodium efflux following the addition of 10 mm-NaCl to FW is reduced by a factor 2·4 (A = 14·4), while the extra sodium efflux due to the total amount of sodium in sea water (550 HIM)-that is the SW flux minus the FW flux-is reduced by a factor 2·2 (A = 12·5). It may be recalled that the effect of the 10 mm potassium contained in the sea water is responsible for only about 4–5 % of the sodium efflux because of competition between the 550 mm-Na and 10 mm-K for the common carrier (see Maetz, 1969). In contrast, the extra sodium efflux due to the addition of 10 mm-KCl to FW is reduced by a factor 5 · 5 which corresponds to the high activation energy of 28 kcal mole⁻¹.

The converse transfer yields for the total sodium efflux a temperature coefficient of 1· 8 (A = 9 · 7), and for the residual flux in FW a ratio of 2 · 3 (A = 13 ·1). The extra sodium efflux due to 550 mm external sodium shows a temperature coefficient of 1 · 7 (A = 8·5). The extra sodium efflux due to the presence of 10 mm-KCl in FW is fully restored immediately upon transfer to higher temperature but a ratio cannot be given as the potassium effect is nil or negative in the cold environment. The FWK flux is, however, increased sixfold (A = 28 · 8). Fig. 5 illustrates comparatively a test performed on a flounder adapted to 6 °C, on the left, and a test performed immediately upon transfer to 16 °C, on the right. Contrary to the experiment illustrated in Fig. 4, the tests were not performed consecutively on the same fish.

In conclusion, the observations reported above demonstrate that the sodium efflux across the gill is a complex process. Some components have a low temperature coefficient: the Na/Na exchange component or extra sodium efflux due to the presence of the external sodium (mean A = 11·3), the residual flux after transfer into FW, an index of the passive electro-chemical flows (mean A = 8·8). The K/Na exchange which presumably corresponds to the active sodium extrusion process exhibits in contrast a very high activation energy (A = 28).

This study is an attempt to gain insight into the branchial mechanisms of osmoregulation in relation to their temperature dependence. Our observations will be discussed from ecological, physiological and biophysical as well as biological viewpoints.

Our results confirm the suggestion made by many authors that a predominant cause of death in fish at low temperature is osmoregulatory failure. This idea was first proposed by Doudoroff (1945). Our data show clearly that it is the exceptional temperature sensitivity of the sodium-extrusion mechanism in contrast to the lesser temperature dependency of the passive transfer mechanisms which is responsible for the disturbance of the mineral balance.

According to a recent review (Umminger, 1969) the mineral imbalance caused by cold adaptation is shown by increased sodium and chloride plasma levels. As may be seen from Table 2, adaptation to a cold environment is accompanied by an increased salt content in the seawater-adapted flounder. Interestingly, however, there is no change in the plasma sodium level. Only the 60% increase of the sodium space detected isotopically is indicative of the perturbation. A survey of the available literature shows that elevated plasma sodium and chloride concentrations do indeed accompany adaptation to cold, but only at sub-freezing temperatures (Woodhead & Woodhead in Gadus morhua, 1959; Pearcy in Pseudopleuronectes, 1961; Gordon, Amdur & Scholander in Myoxocephalus, 1962). In Fundulus heteroclitus, Umminger (1969) finds no difference in serum sodium and chloride levels between fishes adapted to 20, 10 and 4 °C. Only below 4 °C does an increase of the electrolyte level occur. It may well be, however, that mineral imbalance is already present when fishes are progressively adapted from 20 to 10 or 4 °C.

What is the biological significance of the increased sodium space ? As muscle tissue represents an important fraction of the body weight an elevation of the muscle sodium content possibly at the expense of the potassium, a shift consequent upon a failure of cellular sodium extrusion pump is the most likely explanation. Previous observations confirm that the sodium concentration of the muscle fibres is found to be elevated after adaptation to cold seawater in Gadus morhua (Woodhead & Woodhead, 1959) in Cottus scorpius (Eliassen, Leivestad & Moller, 1960) and in various zoarcid and stichaeid fish (Prosser, Mackay & Kato, 1970). A decreased muscle potassium content has only been reported so far for Cottus by Eliassen et al. (1960). In the flounder we find that although the water content is not altered (see Table 7), a reduced sulphate (extracellular) space of the parietal muscle which is indicative of an increased cellular water content. In parallel, the sodium content is increased three-fold, i.e. by 10 μ-equiv ml⁻¹, while the potassium content is decreased by this same amount. Thus partial replacement of potassium by sodium accompanied by fibre swelling is characteristic of cold adaptation. With respect to the extra sodium content of the fish, however, one point has to be made. The total amount of sodium taken up during cold adaptation is about 3500 μ-equiv per 100 g fish (compare in Table 2 the product of sodium concentration and space in fish adapted to 6 °C and to 16 °C). Yet for a muscle-water mass of 50 ml per 100 g fish, the actual increase of sodium content found in muscle amounts to only 500 μ-equiv. Other tissues-the skin, the liver and gut, the skeleton-must participate to the increase in salt content.

The increased salt load related to adaptation to a cold environment was suggested by Gordon et al. (1962) to be due to either an increased water permeability of the gill or to a decreased solute excretion by the gill. The results obtained independently by Motais and Isaia (in preparation) on the effect of temperature on the diffusional water permeability of the gill are not in agreement with the first suggestion, while the present paper concerning the branchial sodium exchange confirms the second hypothesis. As shown in Table 4, after 7–12 days in a cold environment, the Na/K exchange appears to be completely abolished, while the passive sodium exchange is much less affected. These points will be discussed in detail in the paragraph concerning the biophysical impheations of our findings. Table 3 shows that the drinking rate, although considerably reduced at 6 °C, still corresponds to a sodium entry via the gut of about 2530 μ-equiv h⁻¹ (100 g)⁻¹. The cumulative entry by way of the gut could alone account for the 3500 /i-equiv sodium load found within a week of cold exposure and failure of salt-extrusion. One point, however, remains to be investigated. Why does the salt load not increase between 2 weeks and 4 months of cold adaptation? Some compensatory mechanism must develop permitting the fish to remain in steady-state with respect to its sodium balance. The drinking rate may be reduced even more after longer periods of adaptation. After 1 month in 5 °C Motais & Isaia (in preparation) found for the seawater-adapted eel a reduction by a factor 4 when compared to the drinking rate measured in the 15 °C environment. Furthermore, a partial restoration of the extrusion capacity of the gill may occur, an interesting possibility which has not been explored. Histological observations on the cod gill suggest increased activity of the chloride-secreting cells at subfreezing temperature (Fedorov, 1967; Woodhead & Woodhead, ^I965).

Abrupt transfer to high temperatures-for example, 26 °C-also results in prompt osmoregulatory failure as shown by the rapid increase of the plasma sodium level. These experiments were performed before the importance of the Na/K exchange in the sodium extrusion mechanism was appreciated. It would be interesting to verify whether this parameter of sodium efflux is irreversibly depressed by heat exposure. The fact that the total sodium efflux is irreversibly depressed, points strongly in that direction : 26 °C is the upper limit of temperature attained in summer in the Ville-franche aquarium. No flounders survive at this temperature even when the temperature increase is progressive (4–5 months). As a matter of fact, the Atlantic flounder studied here lives on the continental shelf near the mouth of the Loire River. The 16-6 °C margin is very similar to that effectively encountered by the fish in its natural environment. It would be most interesting to make a comparative study of the Mediterranean flounder whose habitat must correspond to a much higher temperature range.

Rapid transfer experiments were devised in order to obtain information concerning the temperature coefficient of the various flux components of the gill. Before discussing the possible biophysical implications of our findings, it is essential to ascertain whether the observed effects result from direct or indirect action of temperature. The gill is a complex organ fulfilling many functions : respiratory gas exchanges, excretion of nitrogenous wastes, maintenance of acid-base and mineral balances. Accordingly the morphology of the gill is complex and the thin respiratory epithelium of the lamellae contrasts with the thick epithelium bordering the filaments between the lamellae (see review by Maetz, 1971). The haemodynamic studies of Steen & Kruysse (1964) suggest that the blood flow may be diverted along two main pathways : the central core of the filaments bordered mostly by mucous and mitochondria-rich cells or the lamellae. The presumably osmoregulatory active cells are the mitochondria-rich cells with a deeply infolded lateral and basal membrane, the so-called ‘chloride-cells’. These cells are found mostly on the afferent side of the central core or at the base of the lamellae. The branchial haemodynamics may be under nervous or endocrine control (Rankin & Maetz, 1971). Catecholamines decrease the resistance to blood flow while acetylcholine and neurohypophysial hormones increase it. Recent observations by Kirschner (1969) indicate that abrupt lowering of temperature results in an increase in gill capillary resistance. A rearrangement of the blood flow between the major pathways may explain why active transport occurring in the chloride-cells is more affected than the passive sodium exchange which probably occurs across the thin respiratory epithelium. The most striking observation by Kirschner (1969) is the 70-fold reduction of the rate of disappearance of radioactive sodium from the external bath into the gill when the temperature of the perfusion fluid is decreased by 15 °C. This effect cannot be ascribed to simple permeability changes of the gill epithelium but rather to changes in the exchange surfaces.

Several points, however, argue against the possibility of temperature effects via haemodynamic rearrangement in intact fish. The major point is whether blood flow may be a limiting factor in branchial sodium exchange. The blood flow in teleosts of the 100-300 g body weight range is of the order of 150 ml h⁻¹ (100 g)⁻¹ (see Discussion by Motais et al. 1969), which brings an hourly supply of at least 15 m-equiv sodium to the gill-that is, 6 times the amount renewed and more than 100 times the quantity excreted. The problem is to recognize whether local rearrangements are able to account for limiting the sodium supply to the branchial epithelium.

The following may be cited as indirect evidence against the possibility of haemodynamic factors intervening in sodium exchange. From the figures given by Kirschner (1969) it can be seen that rapid cooling of the perfusate is followed by a progressive increase in aortic pressure. Up to 1 h is necessary for the increase to be completed. Abrupt temperature drop in our experiments is followed within 5 min by a flux decrease which remains constant during the subsequent 31 min of the test (see Fig. 1). Furthermore, Kirschner (1969) insists upon the relative irreversibility of the gill permeability change after increase from 5 °C to about 20 °C. In our preparation, reversibility of the flux changes prevails between 6 °C and 21 °C. In fact, the excessive leakiness of the artificially perfused gill preparation, prevented at low temperature and favoured by high temperature, probably explains the irreversible change observed as well as the huge flux change in relation to temperature change. In the experiments reported here the flux change is barely more than double between 21 °C and 6 °C.

Hearon (1952) has reviewed the difficulties of interpreting temperature coefficients in complex biological systems and particularly how to identify the importance of the rate-limiting steps or ‘master reactions’. An example of the difficulties encountered in the study of the temperature dependence of active and passive transport is provided by the isolated frog skin or toad bladder. On this biological membrane all studies agree that the active transport as measured by the short-circuit current is positively dependent on temperature up to 25–28 °C and negatively dependent at higher temperature (Snell & Leeman, 1957; Dalton & Snart, 1967 a, b;Porter, 1970). The latter two investigations disagree, however, on the important point of whether the rising phase of temperature-dependence follows or not the Arrhenius formulation. Porter (1970) studied bi-directional sodium fluxes by a double-labelling technique and found that the rising phase of the temperature response of the short-circuit current (i.e. the net sodium flux) is correlated with an increase of sodium flux, the efflux remaining constant. At higher temperature, however, the efflux becomes progressively more temperature-dependent than the influx, hence the decreased net sodium flux. In any case, as pointed out by Porter, for the rising phase of temperature response the sodium influx composed of both passive and active fluxes is more temperature-dependent than the efflux, which is mostly passive, so that it seems clear that the primary effect of temperature is chiefly on the active component and not on passive entry of sodium as claimed by Dalton & Snart (1967 a, b). But even with active sodium transport multiple sites of action of temperature, including direct effects on the sodium pump or on the energy supply to the pump, have to be taken into account. Both these factors intervene as shown by studies involving inhibitors ; both ouabain, an inhibitor of the pump, and metabolic inhibitors impair the response to abrupt temperature change (Porter, 1970). That the temperature-dependence of the sodium transport appears to be a multi-site process is also shown by the progressive increase of the short-circuit current following abrupt temperature change ; about $12$ h is necessary for the full effect of temperature change to be complete.

In our biological material, the available evidence argues in favour of simpler processes of sodium transfer in relation to their respective temperature coefficients. The total sodium efflux, which contains a very small component of active sodium extrusion, follows the Arrhenius relationship between 6 and 21 °C. The Q₁₀ over this range is about 2 and the activation energy 10 kcal.mole⁻¹. As earlier discussed, the sodium efflux probably contains two components : simple diffusion along the electro-chemical gradient and exchange-diffusion. Both these components are not expected to be very temperature-dependent. The sodium leak observed after transfer of the flounder from sea water into fresh water which may be considered as an index of the simple passive diffusion permeability of the gill shows a temperature coefficient of about 2 and an activation energy of 10 kcal.mole⁻¹. Both these parameters have therefore similar activation coefficients which are about twice that measured for free diffusion of sodium ions in water (A = 4·7 kcal. mole⁻¹ ; see review by Stein, 1967). In fact, sodium exchange in highly cross-linked resins (8 to 12 kcal.mole⁻¹, see Stein, 1967) is very similar to sodium exchange across the gill in terms of temperature dependence. It follows that passage through the branchial membrane is certainly not by a process of free-diffusion through wide water-filled pores.

Finally, the potassium-dependent sodium efflux appears to be very highly dependent on temperature changes. An activation energy for sodium transport of 28 kcal. mole⁻¹ has been calculated for this parameter. Similarly high values have been given for the Na/K exchange processes of the human erythrocyte or ascites tumour cells (see review by Stein, 1967). More work is necessary with metabolic or pump inhibitors to decide whether the rate-limiting step is the supply of energy or the activity of the pump. In any event our claim that sodium extrusion in teleostean gills kept in sea water is an active process is substantiated by our observation on its temperature dependence. It may be recalled, however, that some processes assumed to be passive-for example, the transfer of bivalent and trivalent anions across the erythrocyte membrane (Passow, 1969)-exhibit unusually high temperature coefficients (A = 30 kcal. mole⁻¹). Caution is therefore necessary before any firm conclusion may be drawn from our evidence.

Our results concerning long-term adaptation to a cold environment point to a total absence of acclimatization to temperature on the part of the sodium efflux and its components. Changes in the rate functions during temperature acclimatization of poikilotherms have been classified by Precht (1958) into several types. In some instances overcompensation, perfect compensation or partial compensation is observed. Examples of such adaptations concern total metabolism or oxidative enzyme activity of tissues (Hazel & Prosser, 1970). The explanation of this compensation is usually in terms of maintaining relative constancy of energy liberation. In other cases absence of compensation or inverse compensation is observed, the acclimatized rate in the cold being equal or lower than the rate observed upon direct transfer from warm to cold.

Concerning the total sodium efflux, our results suggest absence of compensation (see Tables 2, 4). For the Na/K exchange inverse acclimatization is observed, as upon immediate transfer a small but significant potassium effect is found, while a week or two later external potassium decreases the sodium efflux. It would be of interest to investigate the Na-K-dependent ATPase activity of the gill in relation to cold adaptation. In unpublished observations by K. Kato reported by Hazel & Prosser (1970) an increase of this enzyme activity in the gills of the goldfish in relation to cold adaptation has been found, while the residual (Mg-dependent) ATPase activity on the contrary shows inverse compensation. Motais (1970) has shown, however, that the Na-K-dependent ATPase activity of the gill in relation to life in fresh water may evolve independently of the Na-K-dependent ATPase activity in relation to life in sea water.

When the present investigation was started, we hoped to find compensatory acclimatization in a cold environment and increased transport efficiency by synthesis of more transport carrier. It is of interest to note that, Motais & Isaia (in preparation) describe a partial compensatory acclimatization for the branchial permeability of the eel to water.

1. The effects of short-term and long-term temperature changes on the branchial components of sodium balance have been studied in the seawater-adapted flounder Platichthys flesus.

2. When fish adapted to 6 °C are compared with fish adapted to 16 °C a disturbance of sodium balance is observed ; while the plasma level of sodium remains constant, an increased internal sodium space can be demonstrated isotopically. Increase of the muscle sodium content accounts for only a small part of the extra sodium content of the fish. The increased sodium load is the result of an impairment of the sodium-extrusion mechanism in the gills, and is demonstrated by the disappearance of the Na/K exchange activity of the gills. The passive sodium fluxes (by simple diffusion or exchange-diffusion) decrease only twofold.

3. Abrupt temperature changes in the 6–21 °C temperature range are followed by instantaneous and reversible changes of the total sodium efflux (Q₁₀ = 2), of the sodium leak observed after transfer into fresh water (Q₁₀ = 1·7) and of the Na/Na exchange (Q₁₀ = 2). The Na/K exchange, which corresponds presumably to the active sodium extrusion mechanism, shows in contrast a much greater temperature dependence (Q₁₀ = 6). The total sodium efflux follows the Arrhenius relation between 6 and 21 °C.

4. Abrupt transfer to higher temperatures (23-26 °C) produces irreversible damage to the transport system.

5. No compensatory acclimatization of the flux rates is observed during adaptation in the 6 °C environment.

6. The biological, physiological, ecological as well as biophysical implications of these findings are discussed. Particular emphasis is given to the problem of gill haemo-dynamics.

The authors wish to thank MM. S. Lenkauer and R. Tanguy for their valuable technical assistance. We are also grateful to Dr A. Cuthbert for his critical comments on the manuscript.