Ionic Permeabilities of the Isolated Gill Cuticle of the Shore Crab Carcinus Maenas

Osmoregulation and ionoregulation in decapod crustaceans have been related to water and salt exchanges across the surface of the gut, the antennal gland and the gills. The latter is the main site of passive loss and active uptake of salts in hyperregulators (Mantel & Farmer, 1983). Isolated decapod gills have therefore been used with the aim of analysing Na⁺, NH₄⁺, H⁺, Cl⁻ and HCO₃⁻ transport mechanisms. In relation to these studies, diffusional transgill potentials have been measured in the crayfish Austropotamobius (Croghan, Curra & Lockwood, 1965) and in several crab species: Maja (King & Schoffeniels, 1969), Eriocheir (Pequeux & Gilles, 1981; Gilles & Pequeux, 1985), Callinectes (Mantel, 1967; Smith & Linton, 1971) and Carcinus (King & Schoffeniels, 1969; Siebers et al. 1985; Pequeux, Wanson & Gilles, 1984; Lucu & Siebers, 1986). From these measurements it has been suggested that the gills of crayfish and shore crabs are more permeable to Na⁺ than to Cl⁻.

Any possible action of the cuticle as a barrier (Mantel & Farmer, 1983) has generally been ignored in the studies quoted above on decapod gills. At least two factors have to be considered to relate the transbranchial ion transport to the cuticle properties: first, the cuticular ionic selectivity or relative ionic permeability across the cuticle, and second, the ionic cuticular permeability relative to that of the underlying epithelium. These two factors should be considered in conjunction with the properties of the subcuticular spaces that are limited by the apical infoldings of the gill epithelial membrane. Such spaces were first described by Copeland & Fitzjarrell (1968) in Callinectes, and can be observed in most hyperregulators, including Carcinus (Gilles & Pequeux, 1985; Towle & Kays, 1986).

Avenet & Lignon (1985) observed a high ionic permeability of the gill cuticle to all ionic species (10 ⁻³cms⁻¹) in the osmoconforming lobsters, Homarus and Nephrops. Such a high permeability of the cuticle is unlikely to affect the salt exchanges across the gills. However, in hyperregulating freshwater crayfish, Avenet & Lignon (1985) and Lignon & Lenoir (1985) showed a high ionic selectivity of the isolated gill cuticle. Besides, the permeability of the crayfish gill cuticle computed from conductance measurements was found to range from 10⁻⁷ to 10⁻³cms⁻¹ depending upon the ionic species. These authors pointed out that, in crayfish, the cuticle could impede the transbranchial flux of some ionic species while allowing almost free passage of other ionic species. In their investigations on Callinectes gills, Smith & Linton (1971) also measured the diffusional transcuticular potential and stated that the preferential permeability of the gills to Na⁺ ‘is primarily a function of the cuticle’. Unfortunately, these authors did not make any estimate of the cuticular permeability so their statement is partly speculative. In contrast, Webb (1940) claimed that the permeability of Carcinus gill cuticle ‘is such that it does not affect the salt and water exchanges of the animal’. However, this claim was based on indirect evidence since Webb measured the salt permeabilities of the cuticle of the lobster foregut and presumed that the gill cuticle of Carcinus is more permeable since it is thinner.

In the present investigation, it is shown that the gill cuticle of Carcinus maenas has a marked selectivity for cations over anions. The cationic permeability computed from conductance measurements is relatively high while the anionic permeability is moderate. These results are discussed in relation to the overall permeability of the crab gill.

The experiments were performed on large specimens of Carcinus maenas at intermoult stage at the Marine Biological Station in Roscoff (Brittany, France). The animals were kept in large tanks of running sea water (SW) prior to the experiment.

The isolated posterior gills were perfused with crab saline (CS) containing Evans Blue and, after a wash with pure CS, a lamella was isolated and cut along its edge with fine scissors. The two sides of the lamella cuticle were gently pulled apart from each other. Any patch of epithelium (selectively absorbing the dye) still adhering to the cuticle was peeled off mechanically.

The cuticle was sandwiched vertically between two identical half-chambers as described by Avenet & Lignon (1985). The volume of the half-chamber facing the exposed area of the cuticle (3 ·14 mm²) was 20 mm³. The cuticle was superfused on both sides at a rate of 10 –20 ml min⁻¹ at room temperature (19 –20 °C). Salines flowing by gravity from storage vessels could be changed using multiway taps connected to the bottom of the half-chambers. Excess saline was removed by suction at the top of the chambers.

Experimental fluids were filtered sea water (SW) from the Station, artificial sea water (ASW) and crab saline (CS) made up with distilled water and pm analysis grade salts. ASW composition was close to that of SW, consisting of (in mmol 1⁻¹), Na⁺, 479; K⁺, 10; Ca²⁺, 10; Mg²⁺, 55 ·5; Cl ⁻, 559; SO₄^{2 −}, 29; HCO₃⁻, 2 ·5; pH8 ·1. CS had the following composition (in mmoll ⁻¹): Na⁺, 465; K⁺, 10; Ca²⁺, 10; Mg²⁺, 8; Cl ⁻, 503; SO₄^{2 −}, 4. The pH was adjusted to 7 ·5 with Hepes (2mmoll ⁻¹) and Tris.

In another experimental series, the cuticle was superfused with single salt solutions. The reference solution was usually a 500 mmol I ⁻¹ NaCl solution, having an ionic strength close to that of SW and CS. The reference concentrations for CaCl₂, MgCl₂ and Na₂SO₄ were 200mmoll⁻¹ and 100mmoll⁻¹ for MgSO₄. CaSO₄ was used at 10 mmol 1⁻¹ (imposed by the solubility product). For comparative purposes, MgSO₄ was also used at 10 mmol 1⁻¹. All these solutions were buffered at pH 7 ·0 with the above-quoted buffer. The partial pressure of CO₂ in the gas mixture used to equilibrate the NaHCO₃ solutions was adjusted so as to hold the pH at 8 ·0. At this pH, 97 % of the total CO₂ would be in the form of HCO₃⁻ so that Na⁺ and HCO₃⁻ concentrations can be considered as virtually equal.

Single ion activity coefficients were calculated from the salt activity coefficient given in the tables of Parson (1959) according to the convention of Bates (Meier el al. 1982).

Electrical measurements were performed as described by Avenet & Lignon (1985). The potential (V_oi) was measured with KCI (3 mol l ⁻¹)-agar bridges whose tips were held close to the cuticle. V_oi was measured with respect to the inside. The junctional potentials were calculated using the Henderson equation.

Current-voltage (l/V) curves were obtained using constant-current pulses of varying amplitude fed from a floating current source to the cuticle through two silver plates. The current I_o was taken as positive when flowing inwards. The series resistances (R_s: resistance of the fluid layers separating the tips of the bridges from the cuticle) were also calculated from I/V curves established at the end of the experiment after the cuticle had been disrupted. Means are given with their standard errors.

Ionic flux across a membrane is proportional to the concentration, to the electrochemical gradient and to a permeability coefficient related to the transport properties of the membrane. Simple explicit solutions of the Poisson—Nernst—Planck equations are obtained if the Poisson equation is approximated either by the electroneutrality equation or by the constant-field equation and if boundary conditions are correctly chosen. At zero total transmembrane current (I = 0), the potential (V_O1) is related to the permeability ratio between different ionic species α_ij= P_i/P_j) and to the ionic activities (a_j). Flux rate constants (k_j) and conductances (G_j) from which absolute values of the permeabilities (P_j) can be derived are generally functions of P_j, a_j and V_oi.

When the cuticle was superfused with CS on both sides, the transcuticular potential (V_OI) was equal to zero, as expected from a cell-free system. The potential was negligibly small when SW bathed the outside and CS the inside. The effect of some ionic replacements performed in CS at constant ionic strength are illustrated in Fig. 1. Total replacement of Na⁺ by K⁺ or L1⁺ resulted in changes of V_OI by up to 30mV and indicate the following permeability sequence: K⁺>Na⁺>L1⁺. Ca²⁺ replacement by Na⁺ was almost ineffective. Cl⁻ replacement by CH₃COO ⁻ induced a 4 mV change in V_oi, showing a larger acetate permeability. The effect of the total replacement of Cl⁻ by SO₄²⁻ would support a larger permeability to SCL^{2 −} but in this case [Na⁺]_i was also reduced so that the negative value of V_oi could also indicate a Na⁺ selectivity of the cuticle. Ionic substitutions performed in SW (ASW) had similar effects (Fig. 1), but the changes in V_oi were then of the opposite sign and (slightly larger. These results show that the crab gill cuticle is permeable to monovalent cations.

The small changes in V_O1 observed could result either from a low selectivity among monovalent cations or from a substantial permeability to anions. To distinguish between these two possibilities, saline solutions were diluted either inside or outside. When CS was diluted, V_OI was negative (e.g. —15 mV for a dilution to half concentration; Fig. 2). Conversely, when SW was diluted, V_O1 was positive. As shown in Fig. 2, V_oi increased, respectively, to +16, +49 and +86 mV upon SW dilutions by a factor of 2, 10 or 100. These values of V_oi are close to the monovalent cation equilibrium potentials, showing a high selectivity of the cuticle to cations over anions.

Single salt solutions were used to distinguish the effects of each ionic species. Replacing either CS or SW by a pure NaCl solution (500 mmol 1⁻¹) of similar ionic strength did not change V_OI by more than 0 ·5 mV. Upon dilution of this pure NaCl solution either inside or outside, changes in V_OI were recorded (Fig. 3) that were similar to those obtained with CS and SW dilutions (Fig. 2), clearly showing a selectivity of the cuticle for Na⁺ over Cl⁻.

During dilution of NaCl upon either side, V_oi was close to the equilibrium potential for Na⁺ (E_Na⁺). The initial slopes of the experimental curves (fitted by eye) were 57 ·4 ±0 ·6 and 57 ·2 ± 0 ·9 mV/decade, respectively, for an external and an internal dilution (N=5). These curves were virtually linear over the first two decades but the slopes decreased at higher dilutions so that V_oi tended to a maximum value (140 ± 15 mV for an outside dilution). This reduction of the slope of the curve V_oi/log (a_NaCi) always occurred at higher NaCl activity when the dilution was performed inside. The maximum absolute value of V_O1 was also found to be 20 –50 mV less than for an outside dilution. Similar results were obtained upon dilution of a 500 mmol 1⁻¹ KCl or LiCl solution. Absolute values of V_oi were slightly larger with KCl than with NaCl and slightly smaller with LiCl. It can thus be concluded that Na⁺, K⁺ and L1⁺ are much more permeant than Cl⁻ across the crab gill cuticle.

Similar dilution experiments were performed with 500 mmol 1⁻¹ solutions of Na⁺ salts containing Cl⁻, CH₃COO⁻ or HCO₃⁻. NaCl and NaCH₃COO dilution performed at pH 7 ·0 and 8 ·0 did not show any significant differences. NaHCO₃ dilutions were performed at pH 8 ·0. In every case the cuticle showed a marked Na⁺ selectivity. However, the slopes of the curves V_oi/log(a_Na⁺) increased with the following salt sequence: NaHCO₃, NaCH₃COO, NaCl, indicating a larger impermeability of the cuticle to Cl⁻ than to CH3COO⁻ and HCO₃⁻. For an outside dilution, typical values of the initial slope and V_oi maximum for NaCl, NaCH₃COO and NaHCO₃ were 57 ·5, 57 ·0 and 55 ·5 mV/decade and 150, 120 and 115 mV, respectively.

Dilutions of salts containing divalent ions were also performed to assess the selectivity of the cuticle to Ca²⁺, Mg²⁺ and SO₄²⁻. The slope of V_oi/log (a_Na⁺) was close to 58 mV/decade upon diluting Na₂SC₄ and indicates a much higher permeability of the cuticle to Na⁺ than to SO₄²⁻. V_oi was positive when CaCl₂ was diluted outside. The curve relating V_OI to log (a_Ca²⁺) was linear over two decades and had a slope of 28 ± 1 mV/decade (N = 3). V_oi was slightly negative (—10 mV at most) when MgC1₂ was similarly diluted outside. This indicates that Mg²⁺ and Cl⁻ have similar permeabilities while the cuticle permeability to Ca²⁺ is much higher. The dilution of both CaSO₄ and MgSO₄ (from 10 mmol 1⁻¹ down to 0 ·1 mmol 1⁻¹) outside induced the development of a positive potential, indicating a higher permeability of the cuticle to cations. The slopes of the curves V_oi/log (a_M²⁺) were 28 ± 1 mV/decade for CaSO₄ (N = 4) and 18 ± 4mV/decade for MgSO₄ (N = 5), respectively. Dilutions down to 1/100 were performed from solutions containing either 10 or 100 mmol 1⁻¹ MgSO₄ and gave similar initial slopes. However, the slope of the V_oi/log (aMgsch) curves was reduced when a 100 mmol 1⁻¹ MgSO₄ solution was further diluted to 0 ·1 mmol I ⁻¹. V_OI values of similar amplitude but of opposite sign were obtained when the dilutions were performed inside. These results indicate that SO₄²⁻ is the less permeant ion across the crab cuticle and that Mg²⁺ is the less permeant cation while Ca²⁺ has a much higher permeability.

Permeability ratios between cations were assessed from V_oi measurements in biionic conditions (total replacement of one cation by another one) (Table 1). When 500 mmol 1⁻¹ NaCl was perfused inside and replaced outside, the potential sequence varied according to the sequence NH₄C1, RbCl, CsCl, KCl, NaCl, LiCl, CaC1₂, TrisCi, MgC1₂ for each experiment (as given for the mean in Table 1). The same monovalent cation sequence was also obtained using SO₄²⁻ instead of Cl⁻ as a common ion. The values of V_oi were then slightly larger. Perfusing the cuticle at 10 mmol 1⁻¹ with CaSO₄/MgSO₄ (outside/inside) showed that Ca²⁺ is more permeant than Mg²⁺. Similar information was obtained when MgC1₂ and CaC1₂ were used (200mmol 1⁻¹).

A similar ionic sequence was obtained when replacements of NaCl were performed inside while maintaining a 500 mmol 1⁻¹ NaCl solution outside. The polarity of V_O1 was reversed and the absolute values of V_oi were usually slightly smaller. Exchanging MgC1₂ and CaC1₂ also resulted in a small drop (3 –4 mV) in the absolute value of V_OI. In every case the time course of the change in V_OI was slower when the ionic replacements were performed inside as compared with the effect of an external replacement.

Substitutions of Cl ⁻ in a 500 mmol 1⁻¹ NaCl solution were also performed. In agreement with the much larger permeability of the cuticle to Na⁺, the changes in V_O1 thus observed were small. These substitutions nevertheless confirm the following permeability sequence: HCO₃⁻ > CH₃COO ⁻> CI ⁻.

The area-specific conductance (G) of the cuticle to various salts was estimated from the slope of the 1/V curve obtained when both sides of the cuticle were perfused with identical solutions. Both the total resistance (R_t) and the series resistance (R_s) were measured for each salt and for each concentration. As illustrated on Fig. 4A, the l/V curves were linear showing the ohmic behaviour of the cuticle in these conditions. G was deduced from the difference between R_t and R_s.

The conductance of the cuticle perfused with 500 mmol 1⁻¹ NaCl was G_NaCl = 3 ·4±0 ·7Scm ⁻² (N=5) and was not significantly different from the conductance obtained with CS and SW perfusions. Under these conditions, R_t and R_s were found to be of the same order of magnitude, thus increasing the relative error made for G_NaCl More precise measurements were obtained when the cuticle was perfused with the relatively impermeant salt MgSO₄. In this case R_t was much larger than R_s. was 5 ·2 ± 2 ·7 mScm ⁻² (N = 5) when the cuticle was perfused with 100mmoll⁻¹ MgSO₄ (activity = 15 mmol 1⁻¹). dropped to 3 ·5±0 ·6 mS cm ⁻²(N = 3) when [MgSO₄] was 10 mmol 1⁻¹ (activity = 4mmol 1⁻¹). Replacing Mg²⁺ by Ca²⁺ (CaSO₄ = 10 mmol 1⁻¹) evoked a large increase of the cuticle areaspecific conductance to 84 ·6 ± 38 ·5 mScm ⁻² (N= 3) in agreement with the fact that Ca²⁺ is much more permeant than Mg²⁺.

G_Na⁺ and G_Cl⁻ were also obtained by perfusing both sides of the cuticle with MgSO₄ (100mmoll⁻¹) and progressively replacing either Mg²⁺ by Na ⁺ or SO₄²⁻ by Cl⁻ up to 20 mmol 1⁻¹ (Fig. 4). R, was then kept constant and the change in G coule be directly related to G_Na⁺ and G_Cl⁻ It can be seen that the slope of the l/V curve was almost unchanged when a small amount of SO₄²⁻ was replaced by Cl⁻. In contrast, the slope of the l/V curve increased markedly when a small amount of Mg²⁺ was replaced by Na⁺, indicating a large cuticular G_Na⁺ (Fig. 4A). The change of the slope was linearly related to [Na⁺] (Fig. 4B). This observation allows us to calculate a molar area-specific conductance of the crab gill cuticle to Na⁺ (g_Na⁺). This was found to be 2 ·05 ± 0 ·53 mS cm ⁻² (mmol l⁻¹)⁻¹ (N=5) on the concentration scale and 2 ·96 ± 0 ·76 mS cm ⁻² (mmol 1⁻¹)⁻¹ on the activity scale. The increase in G was noticeably smaller when SO₄²⁻ was replaced by Cl⁻ than when Mg²⁺ was replaced by Na⁺ : 50 ± 5 (N= 3) times smaller.

The experimental results in crab saline or sea water show a high selectivity of the crab gill cuticle for cations over anions normally present in SW and in the haemolymph. This was confirmed in single salt experiments that were performed to obtain precise values of the permeability ratios and of the conductance to each ionic species: monovalent cations and Ca²⁺ are much more permeant than other ionic species. The similarity of the results obtained in both conditions shows that the main properties of the crab gill cuticle are not modified if some ions are excluded from the perfusion medium. This allows for the use of permeability values, determined in single salt experiments, to compute the ionic fluxes in more complex solutions such as SW and CS. These permeability values should then be incorporated into a model that accounts for both the selectivity and the asymmetrical behaviour of the crab gill cuticle.

P_Na⁺/P_anion ratios were deduced from the fitting of the V_oi/log (a_NaX) curves obtained by dilution of Na⁺ salts containing Cl⁻, SO₄²⁻, CH₃COO⁻ or HCO₃⁻. The experimental V_oi/log (a_NaCl) curves cannot be fitted by a model using the electroneutrality condition (equation 2b) that predicts a linear relationship. The curvilinear characteristic of the experimental V_oi/log(a_X) relationships such as those illustrated on Fig. 3 supports of the use of the constant-field condition (equation 2a). A P_Na⁺/P_Cl⁻ ratio of 260 ± 120 (N = 5) gave a relatively good fit (within a few mV) of the V_oi/log (a_NaCl)_o curves. This ratio was reduced by a factor of 2 –3 for NaCH₃COO and NaHCO₃. The fit of the V_oj/log (a_NaCl)_i curves was not as good but supported a P_Na⁺/P_Cl⁻ ratio of about 120. A PM_S²⁺/PSO₄²⁻ ratio of 7 was computed from the V_oi values recorded in diluting MgSO₄. P_Ca⁺/P_SO4²⁻ was found to be greater than 100. The slope of the V_oi/log (a_Ca²⁺) curves (28 mV/decade) also indicated a large P_Ca²⁺/P_Cl⁻ value upon dilution of CaC1₂.

Since the ratios P_cation/P_Cl⁻ and P_cation/Pso₄²⁻ are large, equation 3a was used to calculate the permeability ratios between cations of like valencies from the V_oi values recorded under bi-ionic conditions. The values of the computed ratios are shown in Stable 1 and give the following two sequences: NH₄⁺ (6·7)>Rb⁺ (4·9) > Cs⁺ (4·5) > K⁺ (3·8) > Na⁺ (1) > L1⁺ (0·32) > Tris⁺ (0·032) and Ca²⁺ (12) > Mg²⁺ (1). The values of the permeability of the divalent cations relative to that of Na⁺ were estimated using equation 3b and gave the following sequence: Na⁺ (l)>Ca²⁺ (0·45)>Mg²⁺ (0·043).

The absolute values of the permeabilities were calculated, using equation 4, from the conductance measurements performed at low current intensities when both sides of the cuticle were perfused with identical solutions. The permeability to Na⁺ computed in this way was 7·1 ± l·8×l0^{− 4}cms⁻¹ (N= 5; activity scale) using g_Na+ values obtained by addition of Na⁺ to a MgSO₄ solution. A similar P_Na+ value was deduced from the value of G_NaCl. P_Cl⁻ was, of course, 50 times lower (l·5× 10⁻⁵ ems⁻¹) as ger was less than g_Na+. P_Cl⁻ can also be calculated in a different way from the individual pairs of g_Na⁺ values and P_Na⁺/P_Cl⁻ ratios (as determined above) giving a lower P_Cl⁻ value of 2·9 ± 1·9× 10^{− 6}cms⁻¹(N = 5). The large change in G when MgSO₄ was replaced by CaSO₄ can be entirely attributed to Ca²⁺. P_Ca²⁺ was then calculated as 1·3 ± 0·6×l0^{− 3}cms⁻¹ (N= 3). ⁠, is the sum of G_Mg²⁺ and ⁠. Both permeabilities and conductances were assumed to be proportional to compute P_Mg²⁺ (5 × 10^{− 5}cms⁻¹) and (5 ·7 × 10^{− 6}cms⁻¹).

Permeabilities can be calculated with a limited number of assumptions from conductance measurements performed under symmetrical conditions. However, the permeabilities determined above should be incorporated in a model that accounts for the properties of the cuticle to predict transcuticular fluxes. Apart from its high selectivity to cations, the crab gill cuticle exhibits a functional asymmetry with regard to the time course and absolute values of V_OI recorded when the salines are modified on either sides of the cuticle. This is not predicted by the classical model of Goldman (1943) that makes use of a single homogeneous membrane. The cuticle should, therefore, be divided into at least two functional layers. The larger potential obtained when the salts are diluted outside is best explained by a model such as that developed by Avenet & Lignon (1985) for the Cl⁻-selective gill lamina cuticle of the crayfish.

The model described above thus gives a relatively good account of the properties of the crab gill cuticle as far as Na⁺ and Cl⁻ permeabilities are concerned. The outer selective layer could be tentatively identified as the epicuticle normally facing the SW. Such a conclusion that the impermeability and/or selectivity of the cuticle could be related to the epicuticle has also been stated by Jahn (1936) for the grasshopper egg membrane, by Yonge (1936) for the foregut of the lobster and by Avenet & Lignon (1985) for the crayfish gill lamina cuticle.

Our results on Carcinus maenas gill cuticle confirm and extend those of Smith & Linton (1971) on Callinectes sapidus in that the gill cuticle shows a marked selectivity for cations over anions. The larger V_oi values obtained with Carcinus cuticle are likely to result from technical differences: Smith & Linton only diluted the external saline without perfusing the inside, thus creating an artificial unstirred layer that added to the effect of the endocuticle in reducing the cuticle selectivity. Besides, from our conductance measurements, the permeability of Carcinus cuticle to Ca²⁺ and Na⁺ is about 10^{− 3}cms⁻¹. K⁺ and NH₄⁺ are about five times more permeant, while Mg²⁺ permeability is 5 ×l0^{− 5}cms⁻¹. The cuticle permeability to anions is noticeably lower and lies between 10^{− 6} and 10⁻⁵ cms⁻¹.

It is now of interest to know how far the cuticle could interfere with Na⁺ and Cl⁻ losses and uptake that occur when shore crabs are transferred to media of different salinities. Since Na⁺ and Cl⁻ are found in equivalent amounts in the blood, they should be considered as ultimately being cotransported even though this overall cotransport can to some extent be dissociated when Na⁺/NH₄⁺ and C1⁻ /HCO₃⁻ (Pressley, Graves & Krall, 1981; Henry & Cameron, 1983; Lee & Pritchard, 1985) exchange mechanisms are switched on. The cuticle should not impede Na⁺/NH₄⁺ and Na⁺/Na⁺ exchanges across the gills since its permeability to NH₄⁺ and Na⁺ is high. However, this is not so for a HCO₃⁻ /Cl⁻ exchange, since both Cl⁻ and HCO₃⁻ have permeabilities lower than 10⁻⁵ cms⁻¹. In this regard the cuticle could create a functional subcuticular compartment that should be taken into account when considering such an exchange across the gills. This conclusion could also hold for a Na⁺-Cl⁻ cotransport (either loose or tight) since the salt flux will be limited by the slowly moving Cl⁻ flux across the cuticle to satisfy the zero current condition across it.

These observations are, of course, to be considered in relation to the medium in which the gills are bathed. In full strength SW, it is most likely that the subcuticular ionic composition is close to that of SW. The transcuticular potential will then be zero and the cuticle as a whole will merely act as a single diffusion layer whose equivalent aqueous thickness will be given by the ratio of the cuticle permeability to the diffusion coefficient in water. This equivalent thickness is about 100 μm for cationic exchanges but about 10 mm for exchanges involving anions (either cotransported or counter-transported). For gills perfused with diluted SW, the subcuticular compartment could equilibrate with the external medium. The transepithelial ion leaks would thus be from the haemolymph to the subcuticular compartment. Recycling of these ions by the underlying epithelium could then occur from this compartment, favouring an apparent impermeability of the whole system in the steady state not only to ions but also to water. Alternatively, transbranchial ionic exchanges (either leak or uptake) could occur. In this case, a transcuticular electrochemical gradient will be established that will depend upon the relative permeabilities of the cuticle and of the epithelium and upon the pumping activity of the epithelium.

With respect to the latter alternative, the ionic permeabilities of the isolated cuticle, as determined in the present work, can be compared to the permeabilities of the whole gill (cuticle plus epithelium) deduced from the work of Lucu & Siebers (1986). According to these authors, the unidirectional Na⁺ flux across the isolated gill of Carcinus maenas perfused with 500 mmol 1⁻¹ NaCl on both sides is 4000 μmol 1⁻¹ h ⁻1 and V_O1 is then zero. They also calculated a P_Cl⁻/P_Na⁺ ratio of 0 ·34 for the whole gill. Taking into account a gill surface area of about 50cm²g⁻¹ fresh mass, the overall gill P_Na⁺ and Per would be 5 ×10 ⁻⁵ and 1 ·5 ×10 ^{− 5}cms ⁻¹, respectively. The gill PN_A⁺ is much smaller than the gill cuticle PN_A⁺. However, the overall gill Per is close to the gill cuticle P_Cl⁻. This, therefore, strongly suggests that the effect of the gill cuticle has to be taken into account in the study of transgill ion transport. Along these lines it is worth mentioning that the isolated cuticle and the whole gills both show a Na⁺ selectivity in Callinectes and that for a similar concentration gradient the transgill and the transcuticular potentials are close to each other (Smith & Linton, 1971). Published values of the transbranchial potential in Carcinus maenas (King & Schoffeniels, 1969; Pequeux, Wanson & Gilles, 1984, Siebers et al. 1985; Lucu & Siebers, 1986) also clearly show a Na⁺ selectivity, as we have observed for the isolated cuticle. However, it could be that both the cuticle and the epithelium show a Na⁺ selectivity and more extensive conductance measurements should be made before going further.

Our measurements, therefore, show that the inference made by Webb (1940) should be discarded : the gill cuticle of Carcinus maenas can no longer be considered as freely permeable to all ionic species. As mentioned above, a Na⁺-Cl⁻ cotransport will be limited by the cuticle permeability to the slowest moving ion to satisfy the zero current condition across it. This permeability is 10 ⁻³cms⁻¹ in the lobster gill cuticle. It is less than 10 ⁻⁶cms⁻¹ in the freshwater crayfish (Avenet & Lignon, 1985). It lies between these values for Carcinus maenas gill cuticle. Since the lobster does not osmoregulate while Carcinus maenas has a limited power of regulation and the crayfish has an even higher power of regulation, it is tempting to relate the NaCl permeability of the cuticle to the power of regulation of decapods. A similar observation has been made for the carapace permeability of decapod crustaceans to salt and water: the carapace permeability has been related to the habitat of the animals (Gross, 1957; Herreid, 1969).

The author wishes to thank Professor P. Lasserre, director of the Marine Biological Station in Roscoff, and members of the staff, particularly Drs M. Moreau and J. P. Vilain, for their hospitality and kind material assistance. Mr Corbethau’s skill in designing the perfusion chamber is fully acknowledged.