ABSTRACT
New experimental techniques are described for the investigation of water and electrolyte fluxes in the eel by studying the internal medium, the urine and the external medium. An experimental tank made up of two compartments isolates the water containing the head from the water containing the trunk and tail of the animal. The two water circuits are separated by remote control. Measurement can thus be made without handling the eel previously adapted to experimental conditions.
The freshwater eel shows low branchial exchanges and low chloride urinary losses. A positive correlation between urinary excretion of water and sodium has been shown.
The silver eel’s skin is impermeable to water and chlorides.
The eel reacts to FW-SW transfer by immediately drinking water. The drinking reflex is therefore not triggered by dehydration due to the osmotic gradient.
During SW adaptation the eel presents a transitory hyperactivity phase of the branchial outflux corresponding to plasma hypermineralization.
The eel which has been adapted to sea water for 3 weeks shows the lowest chloride exchanges ever recorded among marine teleosts.
Introduction
The recent articles of Maetz (1971), Lahlou (1970), Mayer (1970) and Motais (1970) have synthesized the present knowledge about the functioning of the peripheral outwardly exchanging organs of osmoregulation in teleosts, namely the gills, alimentary tract and kidneys. The silver eel is very good biological material because of its high degree of euryhalinity, its ability to withstand starvation for a long time and its great physical resistance. It can therefore survive numerous and varied surgical operations very well. However, the numerous studies on the eel have essentially been concerned with sodium exchanges and very little with chloride exchanges. In addition, in the fresh water eel, which is unable to absorb external chlorides, the chloride losses recorded for only one of the peripheral organs would be sufficient to bring about the total chloride depletion of the animal within a few months, whereas the starving silver eel can easily survive for more than a year. It has often been emphasized that water and electrolyte exchanges in the eel are increased by physiological shock, such as the effect of experimental handling on diuresis (Chester Jones, Chan & Rankin, 1969) or on ion exchanges (Mayer & Maetz, 1967), the effect of temperature on chloride excretion (Fontaine & Callamand, 1940), on diuresis and on the drinking rate (Gaitskell & Chester Jones, 1971). These observations underline the absolute necessity of creating non-stressful experimental conditions and of checking that they are physiologically innocuous by long-term experiments. We therefore perfected new experimental techniques adapted to long periods of experimentation before studying water and chloride exchanges in the silver eel. These techniques will later be used to study, in good conditions, the regulatory processes which play a part in the eel’s adaptation to sea-water. For the present, they have shown us that the non-stressed eel is, given our present knowledge, the teleost most isolated from the surrounding medium.
MATERIAL AND METHODS
We shall mention below only the materials and methods used in connexion with those already described in our previous paper (Kirsch, 1972) and which dealt with the following points. Origin of the eels: caught in eel-pots in the Rhine affluents in Alsace. External medium temperature: 13·5 ±0·5 °C. Surgical techniques: anaesthesia (30%o ether), intra-aortic cannulation. Experimental material: description of experimental tank and of its contention apparatus. Plasma analysis : Cl− concentration using a modified Sanderson method, Na+ concentration with a flame photometer, 36Cl− concentration by counting on dry samples or continuously with the modified Istin counting cell.
Surgical techniques: Cannulation of the bladder,
Many different urethral cannulations were first tried. In the best conditions, around 10 days of experimentation can be had with this technique because it causes lesions of the neck and wall of the bladder. Moreover, the ligature of the urethra around the urinary catheter rapidly produces tissue necroses of the papilla.
Insertion via the urethra was avoided by using a flexible polyvinyl cannula made in our laboratory and shown in Fig. 1, which presents the essential steps of the surgical operation. The orifice (B) temporarily made in the body wall to insert the cannula is small and closes shortly after the operation. This technique avoids bladder lesions for two reasons : first, the cannula tube is tightly bound to the musculature of the animal and therefore does not rub against the bladder wall ; and secondly, the position on the cannula of the urine sampling orifices prevents the bladder wall from sticking to them.
It must be noted that in our experimental device urine samples are taken at 5–10 cm of negative water pressure, so that the bladder remains permanently empty and the urinary papilla naturally closed. This makes it possible, with advantage to eliminate the urethral ligature in long experiments, but involves the risk that, if the animal struggles violently, water from the external medium sometimes enters the bladder.
Experimental tank and external water circuits
To the previously described experimental tank was added an intermediate compartment (MbC, Fig. 2) used to separate the water in the anterior part of the tank (AC containing the head) from the water in the posterior part (PC, containing trunk and tail). MbC includes a very flexible rubber sleeve whose diameter can be regulated by adding water into, or taking it out of, the MbC from the MbR container. By regulating the tightness of the sleeve around the eel, complete isolation between AC and PO is assured with a very moderate compression corresponding to 4·5–6·5 % of the diameter of the animal’s body.
The posterior compartment is modified by adding a rubber ballonet (Bp) in the space between the interior polyethylene envelope and the stiff exterior envelope. With this ballonet, the volume of the flexible polyethylene envelope of the PC can be regulated during the measuring sequences and then put back to the original volume determined when the eel is put in the tank.
This experimental tank is connected to the water circuits shown in Fig. 3.
Between measuring sequences the MbC separation membrane is loosened and water circulates freely between AC and PC. The circuits drawn in dots are then functional and the eel is supplied with water from a large reserve bottle integrated into the closed circuit. The water is pumped and oxygenated in an air-bubble circuit (air 1). The eel can also be put in running water connected to the entrances (I) ; in this case, the exits (O) are placed on an exterior drain pipe.
During the measuring sequences the MbC separation membrane is tightened and AC water is completely isolated from PC water. Two special reversing cocks (Rea and Rep) make it possible to isolate, without handling the eel, a reduced-volume closed circuit for AC and another one for PC (drawn in solid lines). Only the AC circuit containing the eel’s head is continuously aerated by an air-bubble circuit (air 2). The circuit contains a ballonet which compensates for the AC volume variations during the measuring sequences. The PC circuit is completely closed and a pump (P) makes the water circulate.
Each circuit contains a side tube for the taking of samples (Sp). To avoid disturbing the graphic recordings of the volumes, a quantity of water equivalent to the volume of the sample is added to the ballonet by a graduated burette (Bura or Burp).
Principle of the graphic recording of volume variations
(a) Maintaining the eel in a stationary position
The changes with time of AC and PC volumes are simultaneously measured. The eel must therefore always remain stationary in the experimental tank. This is achieved through the following regulatory system. A position sensor made up of a small electric light bulb and two photodiodes is placed on either side of the eel’s lower lip (Pos. ca., Fig. 2). If the eel moves forward or backward the sensor transmits the information to an electronic controlling device (Pos. elec., Fig. 3). This, in turn, sets in action the peristaltic pump (PP) which regulates the volume of the ballonet (Bp) and, thus, the volume of the posterior compartment to push the eel forward or draw it back into its original position.
(b) Recording the water-volume variations in AC and PC
The mechanism keeping the eel in a stationary position is used to record the volume variations in PC. Any water volume variation (ΔV) in the PC closed circuit makes the eel move and a movement of the same quantity ΔV of water in the peristaltic pump (PP) brings the eel back to its normal position. ΔV is continuously recorded on a cylinder of smoked paper by a device shown in Fig. 4.
A water-level sensor (Lc) detects any modification in water volume in the recording circuit (Rec. c., Figs. 3, 4). The information from the Lc is transmitted to an electronic recording device which moves the piston of a water and mercury column. When the piston moves, it brings the water in the Lc back to its original level and moves a writing stylet proportionately to the quantity of water supplied from the column. The recording apparatus includes a device for re-setting the column, either as a function of time or by end-course contacts ; during the re-setting, the column is disconnected from the Lc and joined up to an annex container through the action of the electric valves EV1 and EV2.
An identical device is used to record the water-volume variations in the anterior compartment. The water level in the circuit is kept constant by the ballonet (Ba, Fig. 3) linked directly to the Lc of a recording device. The sensitivity of the record under experimental conditions is ±o·1ml for a 10 ml sensor.
(c) Sampling and volume recording of urine
The volume of urine excreted is continuously recorded by application of the same principles as used in the recording of the water volumes in the compartments. A cyclohexane trap is placed between the urinary duct and the Lc recording sensor (Fig. 5). This trap makes it possible to accumulate the urine and to take samplings morning and evening. In addition, a magnet transmits regular pulsations to the urinary duct to avoid the sedimentation of the small amount of solid matter in the urine and the consequent obstruction of the catheter during long experiments.
The recording sensitivity for a 1 ml sensor is ± 0·01ml.
Analysis and counting techniques
The 36Cl-concentration is continuously measured by counting cells inserted in the circuits of the anterior and posterior compartments (CC, Fig. 3).
The quantity of chloride is determined using the modified Sanderson method on samples of about 1 ml of water or urine (precision: 0·07%) and of 50 μA of sea water (precision 0·6%). The quantity of sodium is determined by flame photometry (precision 1 %).
Experimental methods
(a) Pre-adaptation of the eel to experimental conditions
The eel is placed in the experimental tank and the volumes of the anterior and posterior compartments are determined with the separation membrane tightened. The ballonets Ba and Bp are empty. The separation membrane is then slightly loosened until it is about 1 mm from the body wall. This limited communication between the compartments makes it possible to place the eel in the experimental position by regulating the water pressure in PC, while the pressure in AC remains constant.
During the pre-adaptation period the position sensor acts on the electronic device which controls the regulator (P. reg., Fig. 3). Depending on whether the regulator opens or closes, the water pressure in the PC is reduced or increased and the eel is drawn back or pushed forward. The animal is pre-adapted in this way to experimental conditions for at least 2 days before the measurements begin.
(b) Measuring sequences
The study of PC in a closed-circuit involves a mean volume of around 500 ml per eel kg. The length of the sequences is limited here only by the risk of cutaneous necroses at the level of the separation membrane between AC and PC. No lesions were observed in experiments lasting less than five consecutive days.
The minimum volume in AC in a closed circuit is 150 ml per eel kg. The length of the experimental sequences is limited here by the rapid pollution of the water; an initial volume of 40 ml of water per hour of experiment and per eel kg is necessary to assure good physiological conditions.
Having fulfilled this condition, we did not observe any significant difference in the results obtained during experimental sequences varying from 2–15 h with eels in equilibrium with the external medium.
(c) The steps in fresh-water to sea-water transfer
In order not to stress the eels by handling the experimental tank, the fresh water was eliminated by rinsing the tank for 15 min with sea water in an open circuit and without modifying the water pressure. The measuring sequences could not therefore begin until after this short initial period in sea water. We arbitrarily took as time of adaptation the end of the rinsing period.
Theoretical considerations
(a)Nomenclature
The following nomenclature will be used in our paper:
To these abbreviations will be added an index denoting the organ studied:
In accord with the usual practices, Fnet will be positive for a gain by the eel, negative for a loss. Fout and Fin will be given in absolute values.
For the entry of water and electrolytes corresponding to the drinking water, the abbreviation EnD will be used. FinD will be reserved for the water and chlorides which reach the internal medium through digestive absorption. ExD will represent faecal losses.
In addition, we shall often use the following abbreviations :
(b) Study of the exchanges by measurement made on the eel
(c) Study of the exchanges by measurements taken on the eel and on the external medium
Since the parameters V, C, *C and RAS can be measured on different media, the following indices have been used to distinguish them :
36Cl− is injected into the eel, and since the external compartments are limited in volume, the radioactive influx must be taken into account. The following formulae are used.
Mathematical presentation of the data
All the data of the experimental series are noted with the mean ± standard error. The linear regression slope is noted with the mean ± standard deviation and the ordinate at origin with the mean ± standard error reduced by linear regression.
EXPERIMENTAL RESULTS
Urinary excretion
Validity of the bladder cannulation technique
The continuous sampling of the urine by cannulation of the bladder can modify urinary excretion. In fact, the retention of the urine in the bladder is eliminated and may modify the possible water and electrolyte exchanges through the bladder wall (Lahlou, 1970). On the other hand, the mechanical action of the cannula can affect the bladder wall. The physiological value of this technique was therefore tested in two experiments.
- The urinary excretion of water and chloride of one fresh-water eel was first measured before cannulation (analysis of the volume and concentration variations of the posterior experimental compartment, ΔPC), then 1 week after cannulation (direct analysis of the urine, U). The results were obtained over a period of 24 h : In the eel with cannulated bladder, water and chloride loses in the PC are slight and the identity of ΔPC and U thus indicate that cannulation does not greatly disturb the urinary excretion of water and chlorides.
The urinary excretion of two silver eels of approximately the same body weight was simultaneously studied over a long period of time (42 days). The animals were placed parallel to each other in the same external conditions (well-aerated running tap water, 13 °C temperature). The results (Fig. 6) proved that cannulation did not progressively impair urinary excretion; and, furthermore, the autopsy of the eels revealed no lesions of the bladder.
Urinary excretion and external conditions
A rise in temperature generally increases the urinary excretion of water (Mackay & Beatty, 1968) and electrolytes (Pora & Prekup, 1960) in fishes. Handling the animal before measuring will produce the same effect (Pitts, 1934). Accordingly the temperature was kept constant and the period of initial shock (12–72 h after the operation) was eliminated from our experiments.
Unfortunately we had to study the effect of water oxygenation upon urinary excretion. If the tap water in our laboratory is not well aerated it causes a nycthéméral rhythm of diuresis. The latter includes a strong night antidiuresis linked to a drop in the partial oxygen pressure in the tap water (Fig. 7). This antidiuresis was eliminated by aerating the water thoroughly. Complementary experiments showed that an oxygen partial pressure of 70–80 mmHg was necessary to assure normal diuresis in the eel. Yet, when one group of eels was studied in untreated tap water and another in aerated tap water both showed an identical urinary excretion (1·26 ±0·10 ml h−1 kg−1, n = 16 and 1·29 ± 0·12 ml h−1 kg−1, n = 14 respectively). The night antidiuresis was therefore compensated for in the first group by an increased day diuresis.
Moreover, the lower graph of Fig. 7 illustrates how extremely sensitive diuresis is to modifications in the external conditions. At about 8 h an antidiuresis followed by a compensatory diuresis appeared. The antidiuresis was due to the resumption of laboratory activity, which was detected by the eel despite all the precautions taken to isolate the experimental tank. This stress reaction occurred frequently and, unfortunately, could not be avoided. At about 10 h an identical stress reaction appeared due to an accidental drop in the oxygen partial pressure of the water, although the critical level of 80 mmHg was not reached. Only the suddenness of the variation in the oxygen concentration could thus be responsible for the observed reaction.
The effect of external conditions on diuresis also appeared in the experiment presented in Fig. 6. Both eels showed a very clear parallelism in their urinary flow caused by identical reactions to unknown modifications in the surrounding medium.
Urinary excretion in fresh water (Table 1)
A minimum of four consecutive measurements over 48 h were taken on each eel without any other experimental manipulation. The individual results which were used to calculate the means presented in Table 1 showed a parallelism between the quantitative urine excretion of water (Uv) and of sodium (UNa). This parallelism also appears in the graphs and results recorded in Fig. 6. To test the theory of a dependence between the quantitative excretion of water and electrolytes we analysed the correlation between 1/Uv and UNa or between Uv and UO1, and to test the opposite theory, the correlation between i/Uy and CNa or i/Uy and CO1. The correlation probability between the variables (Table 1 b) indicated that the more intense the diuresis the more sodium the eel excretes, and that the chloride concentration in the urine varies inversely with the urine flow so that chloride excretion is independent of water excretion.
However, the linear regression coefficients were subject to very large errors, and, to a large extent, independent fluctuations of each of the parameters were therefore added to the correlations presented.
Urinary excretion during the fresh-water to sea-water transfer
After the FW-SW transfer we found the same time course of adaptation as described by Chester Jones et al. (1969) in the eel. Urine flow decreases very fast, and in less than 6 h of adaptation, reaches a steady level of approximately a tenth of the fresh-water level. Table 2 shows the rates of water, chloride and sodium excretion in the eel, in SW over a period of 24 h in relation to excretion in FW. Owing to unfavourable technical contingencies we have only very little information on sodium excretion, and the individual values observed have been recorded.
Chloride excretion is not affected by the decrease in urine flow, whereas sodium excretion diminishes markedly. This confirms the quantitative relationship between water and sodium put forward above.
Exchanges of water and chloride in the gills and in the alimentary tract
Fresh-water eel
(a) Study of water and chloride fluxes by the variations in the composition of the external medium
The volume and composition of water in the anterior compartment, AC (branchial exchanges and digestive entry), and in the posterior compartment, PC (faecal losses and exchanges in the skin), were studied separately in eels which had undergone cannulation. In order to check that the separation membrane would not cause shock to the animal, preliminary experiments were performed and are shown in Fig. 8. In the first sequence the eel was placed in a single water circuit (separation membrane untightened), and in the second sequence the volume of AC and PC were recorded separately (separation membrane tightened). Diuresis was identical in both sequences and the eel was therefore not ‘stressed’ by the tightening of the separation membrane between AC and PC. All the water absorbed in the anterior compartment was found in the urine, whereas there was practically no change in volume in the posterior compartment. Twenty-five experiments were performed on 11 eels having a body weight of W = 0·586 ± 0·053 kg. The animals were preliminarily marked with 36Cl− (about 20 μC kg−1) so that the following unidirectional and net fluxes in the anterior compartment could be measured:
In the posterior compartment the exchanges were very slight, and measuring sequences of at least 24 consecutive hours were required. In five eels (W = 0·550 ± 0·047 kg) we observed a mean reduction in the PC volume of 0-058 ± 0-049 h−1 kg−1
If the eel was marked with 8#Cl−, the PC radioactivity showed a very irregular alternation of stability phases and increase phases. The eel therefore lost chlorides intermittently. These losses can only correspond to faecal matter which is discharged at irregular intervals. The existence of sequences without measurable losses of 36Cl− means that eel skin is practically impermeable to chloride. In a 0·700 kg eel the chloride losses in the PC, measured morning and evening for 4 consecutive days, showed the following successive values: 0·6, 0·2, 1·2, 1·1, 0·8, 0, 1·5 μ-equiv h−1 kg−1.
(b) Measurement of chloride flux by the study of urinary excretion and plasma composition
The branchial and digestive chloride fluxes are so slight in the fresh-water eel that it is impossible to determine by the preceding technique whether a chloride influx exists. For this reason chloride exchanges in one eel were studied over a 3-month period to obtain a noticeable modification in the plasma chloride. The eel was supplied with a bladder cannula and an intra-aortic cannula making regular blood samplings possible.
Adaptation to sea water
(a) Study of the fluxes by the variations in composition of the external medium
Fig. 9 shows an example of the time course of the experimental compartment volumes, of the urine flow and of the chloride outfluxes after the fresh-water to sea-water transfer. The following observations were made.
After the FW-SW transfer the eel drinks 5–6 ml of sea water, which produces a reduction in AC volume and a parallel increase in PC volume. During the first 6 h after transfer this change continues more slowly and the gills ‘osmotic loss of water is therefore less than the drinking rate. These two factors then become practically equal and PC volume remains stable. The urinary excretion of water is compensated for by a slight reduction in AC volume and thus water absorption is slightly greater than branchial osmotic losses.
The increase in branchial chloride outflux (FoutB) occurs only 12 h after the FW-SW transfer;
The chloride outflux in the PC is large during the first 12 h of adaptation and shows successive peaks; it therefore corresponds to faecal discharge. This emphasize the importance of chloride digestive transit in the initial phase of adaptation and the advantage of an experimental separation of FoutB and FoutD. The volume of external water in the single closed circuit is important and so the distribution of faecal chlorides is slowed and the FoutD peaks are integrated into a continuous change. During the preliminary experiments this led us to refer all the chloride losses to FoutB and, wrongly, to consider the hypothesis of an immediate regulation of FoutB after the FW−SW transfer;
the skin is practically impermeable to water and chlorides because the volume and 36Cl− concentration of the PC are stable after 24 h of adaptation.
Table 4 shows the outflux and net flux measurements taken on AC during short measuring sequences; only a very small amount of water must be used for these measurements so that the concentration variations caused by the exchanges are noticeable. But this makes frequent water changes necessary to avoid artifacts caused by the rapid pollution of the water.
The total quantity of chlorides absorbed by the gills and alimentary tract (FnetB + EnD) decreases very quickly after the FW-SW transfer. We can try to estimate the FinB maximum part in this overall measurement by assuming that branchial permeability to water is null. The maximum values possible for FinB can then be calculated. These calculated values are extremely exaggerated because of the indeterminate osmotic losses of water in the gills; FinB = 0 would correspond here to a branchial osmotic permeability of 0·52 ml h−1 kg−1, which is slight. FinB and FoutB develop at the same rate and the massive chloride absorption during the first hours of seawater adaptation therefore correspond essentially to the water drunk.
Fig. 10 shows the water and chloride exchanges measured in the anterior compartment of seven eels. The ingestion of 5–10 ml of sea water per kg body weight immediately after the FW-SW transfer brings about a massive entry of chlorides into the animal.
The chloride branchial outflux varies considerably from one eel to another both in its time course and in its quantitative level. In fact, the activation delay of the branchial chloride excretion pump varies from 2 to 14 h after the FW-SW transfer (17 eels studied) and the rapidity of the FoutB increase is very variable. Fig. 11 presents the correlative evolution of FoutB and plasma chloride in two eels of approximately the same weight. In eel no. 1 plasma chloride quickly reaches its level of equilibrium, and though FoutB increases very fast it shows a peak of little importance. Inversely, in eel no. 2 plasma chloride develops slowly (hypermineralization phase; Kirsch, 1972), and though FoutB increases slowly and late it shows a very important peak. The quick adaptation of the eel to sea water therefore appears to be closely correlated with its ability to regulate FoutB rapidly.
(b) Study of the chloride exchanges fluxes by the variations in the composition of the internal medium
The eels are adapted to sea water in a closed circuit of 100 1, continuously filtered and aerated. The size of the external chloride compartment makes the return radioactive flux negligible during the radioactive discharge (animals marked with 36Cl− before the FW-FS transfer). The time course, in four eels, of the net fluxes and chloride overall outfluxes are shown in Fig. 12. To represent FinG by the distance between curves FnetG and Fouta, FoutG has been shown in negative values. The overall net flux decreases quickly and becomes zero within approximately 50 h of adaptation; this agrees with our previous observations (Fig. 10). On the other hand, the absolute value of FnetG immediately after the FW-SW transfer is only 1000 μ-equiv h−1 kg−1 as compared to 5000μ-equiv h−1 kg−1 for FinB + EnD measured in the external medium (Figs. 10, 12): this is accounted for by the time lapse between the ingestion of the chlorides with the drinking water and their appearance in the internal medium by intestinal absorption. Like the branchial outflux, the chloride overall outflux differs greatly from one eel to another both in its rapidity of increase and in its activation delay after the FW-SW transfer.
Chloride turnover rate in eels adapted to sea water
In eels which have been adapted to sea water for 3 weeks FnetG is zero and FinG is analysed through the simultaneous measurements of chloride space (E), plasma
DISCUSSION
Urinary excretion of water and electrolytes
The eel’s urinary excretion has already been studied under various experimental conditions and at various temperatures (Smith, 1929; Sharratt, Chester Jones & Bellamy, 1964 a; Butler, 1966; Chester Jones et al. 1969; Gaitskell & Chester Jones, 1971). For the excretion of water and sodium, our observations on the fresh-water eel agree with those of Chester Jones, which were made under similar experimental conditions. We have found, however, a quantitative excretion of chlorides twice as small (I·66±0·34 instead of 3·62 ±0·42μ-equiv h−1 kg−1). The kidney-bladder system in the fresh-water eel therefore possesses an extremely efficacous capacity for chloride retention as compared with that of other freshwater fishes (Lahlou, 1970). This adaptation is particularly useful because the eel is incapable of compensating for its urinary losses by a branchial active absorption of chlorides in fresh water (Krogh, 1939; Garcia Romeu & Motais, 1966).
The comparison of the results reported by the different authors mentioned shows, moreover, that eels which pass the most urine in fresh water lose the most sodium. An identical correlation can be noted issssn the observations of Maetz et al. (1964) on Carassius subjected to the action of the hypophysial peptides and in those of Butler (1966) on the normal and hypophysectomized eel. The positive correlation between water and sodium was also found in our freshwater experiments, both in comparing different eels and in comparing successive sequences of excretion in a single eel. On the other hand, chloride excretion is independent of water excretion, and chloride concentration diminishes when diuresis increases ; this may explain why Sharratt et al. (1964 a) did not find any chlorides in the very abundant urine of eels suffering from shock in thermal diuresis.
These correlations still appear after the FW-SW transfer; during the early hours of adaptation the eel presents a parallel reduction in the quantities of urinary water and sodium, while quantitative chloride excretion is not affected at all.
The fresh-water eel’s diuresis revealed the eel’s extreme sensitivity to harassment and particularly to oxygenation changes in the external medium. It is known that the eel shows an increased and prolonged diuresis of post-operative stress associated with an increase in the drinking rate (Gaitskell & Chester Jones, 1971). In addition, the eel shows antidiuretic reactions to lack of oxygen or to brief harassment. These reactions are followed by a compensating diuresis and are probably linked to a modification of the circulatory dynamics and, consequently, of the glomerular filtration rate ; Labat, Peyraud & Serfaty (1962) showed, in fact, that brief harassment (light or approach stress) bring about a transitory bradycardia in marine teleosts.
A long study of urinary excretion is therefore an excellent test for checking the physiological value of new experimental conditions, and we have retained only methods and techniques which do not disturb the diuresis of reference eels.
Extra-renal exchanges of water and chlorides
(a) Impermeability of eel skin to water and chlorides
The absence of a radioactive outflux in the posterior compartment (PC) during certain experimental sequences shows clearly that eel skin is impermeable to chlorides. On the other hand, PC volume represents the sum of the faecal losses of water and of the osmotic net flux of water in the skin. The stability of PC volume in fresh water could therefore result from a balance between these two parameters. Yet in these conditions the reversal of the osmotic flux of water in SW should produce a very important lack of balance. But PC presents a stable volume in SW. The osmotic flux of water and the faecal losses are therefore negligible.
This impermeability of eel skin is parallel to the one observed, in vitro, by Fromm (1968) on the skin exchanges of water and sodium in Salmo gairdneri. The author assigns this impermeability to the poor cutaneous vascularization of teleosts and to the presence of scales and mucus. The thickness of the skin may also play a part ; the eel has, for example, a skin 7 times thicker than the frog’s and much less vascularized (Jakubowski, 1960). It would be interesting to check whether the yellow eel, with its thinner skin, presents the same impermeability as the silver eel.
(b) Branchial and digestive exchange fluxes in fresh water
In the fresh-water eel we have confirmed the functional inability of the gills to bring about positive chloride exchanges, but our measurings of the fluxes are lower (FoutB : 5–20 times lower) than all previous results, in vivo by Motais (1967), Maetz (1971) or in vitro by Kirschner (1969). Because the gills are only slightly permeable, the eel can limit the loss of chloride from its internal medium. This also explains its ability to endure starvation for long periods. Two facts may explain the very low level of chloride exchanges measured: (i) the reduction to the strict minimum of stress phenomena which increase branchial ion exchanges (Maetz et al 1964, Mayer & Maetz, 1967) and (ii) the large size of the eels studied; Utida et al. (1966) have shown, in fact, that the yellow eel’s gills are more permeable to sodium than those of the silver eel.
We were not able to measure directly the part played by digestive absorption in the influx (FinG). But this part is certainly small because the quantity of drinking water was necessarily smaller than the urine excreted by our animals. FinG must, therefore, be essentially branchial (FinB) and the ratio between FinB and FoutB is very much the same in our measurements and in those made by Kirschner (1969) in vitro for a level of chloride exchanges 20 times higher. The eel can therefore reduce its branchial exchanges of chlorides to a very low functional level without modifying the mechanism of unidirectional fluxes.
(c) Branchial and digestive exchange fluxes in sea water
The chloride fluxes we measured in the eel adapted to sea water represent the lowest values observed till now in marine teleosts. Most studies of ion exchanges deal, however, with sodium exchanges, and a quantitative comparison can therefore only b| made if sodium and chloride fluxes are equal, which is precisely the case in the eel (Motais, 1967; Maetz, 1970).
Among the marine teleosts studied up till now, Opsanus tau is nearest to our eels in its low rate of peripheral ion exchanges. But it must still be noted that the mean turnover rate (λ = 6·9 %) we measured in the eel is lower than the minimum turnover rate (λ = 7·5%) recorded by Lahlou & Sawyer (1969) for Opsanus tau. Just as in fresh water, the eel in sea water is therefore markedly isolated from the external medium, and it also saves on the energy needed for active ion transport.
During the FW-SW transfer we observed that the eel reacts drastically to a change in medium by drinking a large quantity of sea water (5–10 ml per kg during the first hour). Taking the change of body weight as a basis, it has been held up till now that the ingestion of sea water is secondary to the dehydration of the eel during the first hours of adaptation (Keys, 1933; Sharratt, Bellamy & Chester Jones, 1964b; Oide & Utida, 1968). In fact, the eel fills up its alimentary tract from the moment it arrives in sea water and even goes beyond its intestinal absorption capacities since a large quantity of chlorides is ejected through the rectum during the first hours of adaptation. Later on faecal losses of chlorides become negligible; this corresponds to the structural modification of the intestine and to the increase in its absorption ability (Hirano, 1967; Hirano & Utida, 1968; Oide & Utida, 1967; Skadhauge, 1969).
The drinking reflex occurs too early for it to be a result of dehydration. It constitutes either a reaction to osmotic shock or a reaction to a specific appetite such as that described in Carassius by Copans & Mayer (1967).
During SW adaptation our eels showed a time course of the branchial outflux (FoutB) which was very different from the one described by Motais (1967) and by Mayer & Nibelle (1970). These authors observed a progressive increase in FoutB to the level of adaptation characteristic of sea-water life (for Motais 13 000 and for Mayer & Nibelle 7000 μ-equiv h−1 kg−1 respectively).
We observed a peak-shaped time course for the branchial outflux, which presents a hyperactivity phase and a secondary reduction down to about 2500 μ-equiv h−1 kg−1 in the adapted eel. These changes show an interesting parallel with the peak-shaped time course of the drinking rate during adaptation in the Japanese eel (Oide & Utida, 1968). A similar temporary increase in drinking rate above the normal level in Anguilla anguilla could explain the plasma hypermineralization phase (Kirsch, 1972) and the peak-shaped time course of the branchial outflux. This hypothesis involves a close correlation between digestive ingestion of sea water and branchial ion excretion. It has already been shown by Mayer & Nibelle (1970) that in the eel in sea water FoutB is regulated by the ion charging of the internal medium. The respective roles of the alimentary tract and the gills in plasma hypermineralization still remain to be studied and are the object of our current research.
SUMMARY
New experimental techniques are described for the investigation of water and electrolyte fluxes in the eel by studying the internal medium, the urine and the external medium. An experimental tank made up of two compartments isolates the water containing the head from the water containing the trunk and tail of the animal. The two water circuits are separated by remote control. Measurement can thus be made without handling the eel previously adapted to experimental conditions.
The freshwater eel shows low branchial exchanges and low chloride urinary losses. A positive correlation between urinary excretion of water and sodium has been shown.
The silver eel’s skin is impermeable to water and chlorides.
The eel reacts to FW-SW transfer by immediately drinking water. The drinking reflex is therefore not triggered by dehydration due to the osmotic gradient.
During SW adaptation the eel presents a transitory hyperactivity phase of the branchial outflux corresponding to plasma hypermineralization.
The eel which has been adapted to sea water for 3 weeks shows the lowest chloride exchanges ever recorded among marine teleosts.