Magnesium Transport by the Urinary Bladder of the Crab, Cancer Magister

Crustaceans successfully inhabit marine, estuarine, freshwater and terrestrial environments. The volume and ionic composition of the internal medium in higher (Decapod) crustaceans are, in part, regulated by the antennal glands (Robertson, 1960; Lockwood, 1962; Potts & Parry, 1964; Riegel, 1972). In marine Decapods the antennal gland is thought to be involved in control of haemolymph volume, hyporegulation of magnesium and sulphate in the haemolymph, excretion of organic compounds and reabsorption of fluid, sugars and amino acids from the urine (Riegel, 1972; Riegel & Cook, 1975).

Although it seems likely that the antennal gland functions in magnesium hyporegulation in marine crustaceans by producing a magnesium-rich urine (Riegel & Cook, 1975), the antennal gland has not been definitively shown to be the site of this hyporegulatory ability. In the marine crab, Pachygrapsus crassipes, the terminal element of the antennal gland, the bladder, has been implicated as a site of magnesium secretion into the urine (Gross & Capen, 1966). Franklin, Teinsongrusme & Lockwood (1978) have shown that magnesium secretion, but not fluid reabsorption, in the antennal gland of the prawn, Palaemon serratus, is apparently controlled by a mechanism located in the eyestalk. The present study was undertaken to determine if the antennal gland is responsible for magnesium hyporegulation in the crab, Cancer magister, and, if so, to determine the role of the bladder and the nature of the control mechanism. Some of the data have been previously published in abstract form (Holliday, 1978 a).

Male, intermoult Cancer magister, weighing 200-800 g, were collected in Coos Bay, Oregon, U.S.A., between June 1975 and August 1977. They were transported to the Oregon Institute of Marine Biology, Charleston, Oregon, and maintained in running Coos Bay sea water (30-34 ‰, 10-14 °C) as previously described (Holliday, 1978b). Crabs were acclimated to 100% sea water (33-34‰salinity, natural sea water adjusted with Instant Ocean ® sea salts, Aquarium Systems, Inc., Eastlake, Ohio) or to 75% sea water (24‰) for 48 h prior to use in experiments. Preliminary studies indicated that urinary rate and serum and urine magnesium concentrations were stable after 36 h in these media. Haemolymph was sampled with a hypodermic syringe inserted at the arthrodial membrane near the base of a walking leg, allowed to coagulate in a small glass vial, and serum was withdrawn from the clot. Urine was sampled at the nephropore by lifting the urinary operculum with a fine, hooked needle and collecting the urine with a pipette. Serum and urine samples were stored in sealed glass vials at — 10 °C until ion analyses were performed.

The effect of low serum magnesium levels (magnesium depletion) on magnesium levels in the urine was assessed by maintaining crabs in magnesium-free 75 % sea water (salinity 24‰). Artificial sea water (34‰) was made from dry salts (Welsh, Smith & Kammer, 1968), with the exceptions that additional NaCl was substituted for MgCl₂ and the trace elements were omitted. The resulting magnesium-free sea water was diluted to an equivalent salinity of 25‰. Crabs, were held in this medium for 6 days, during which time the medium was changed twice. Serum and urine samples were taken at the end of this period. Crabs showed no signs of distress after this treatment, but they did appear to be easily excited and aggressive.

Because crabs void urine at a low rate and at irregular intervals in 100% sea water (Hunter, 1973), experiments to determine the role of the antennal gland in eliminating injected MgCl₂ (magnesium loading) were conducted with crabs acclimated to 75% sea water. The crabs were fitted with rubber tubes for measurement of urinary rate by external collection of voided urine (Holliday, 1977) and placed in individual aquaria supplied with recirculated 75 % sea water. After 48 h acclimation, serum magnesium levels were approximately doubled by injection of 0.12% body weight (ml/g) of a 3.26 M solution of MgCl₈ through the arthrodial membrane at the base of a walking leg. For a 475 g crab with ≃120 ml haemolymph, this represents an additional 3.7 m-equiv Mg per crab. Serum and urine samples were taken and urinary rate was measured over the 48 h periods preceding and following magnesium loading. The last 12 h of acclimation to 75 % sea water (36-48 h after the start of the experiment) was used as a control period for comparison with urinary magnesium excretion after magnesium loading. Urinary magnesium excretion was calculated as urine volume x urine magnesium concentration and is expressed as m-equiv Mg excreted/day.

Experiments to determine the role of the antennal gland in magnesium hyporegulation were conducted with anuric crabs. Large crabs weighing 600-800 g were secured dorsal side up in a dissecting tray and anaesthesia was accomplished by covering the crabs with ice for 20-30 min. The area of the dorsal carapace behind each eyestalk was carefully cleaned with acetone and a 1 cm diameter circle of hot-melt glue (Thermogrip ®, U.S.M., Reading, Pa.) was built up on the carapace just posterior to each eyestalk as a quick succession of thin spots so as not to burn the hypodermis beneath the carapace. The circles of hot-melt glue were lifted free of the carapace one side at a time and glued to it with α-cyanoacrylate glue (Krazy Glue, Inc., Chicago, Ill.). The area within the circular beads of glue was then cut out with a high speed drill, and the wound was slowly irrigated with crab Ringer solution (CR, containing in m-equiv/1: 449 Na, 11 K, 25 Ca, 37 Mg, 528 Cl and 44 SO₄ at pH 7.8) to wash away excess haemolymph and to prevent entry of air into the haemocoele. With the aid of a dissecting microscope, the hypodermis was removed and the bladder was retracted to expose the left antennal artery. The coelomosac artery was located and cut near its point of exit from the antennal artery and the wound was carefully sealed with hot-melt glue in such a manner that the exposed tissues were not heated. The operation was then repeated on the crab’s right side and the crab was returned to 100% sea water. Control crabs were treated similarly, but the coelomosac arteries were not cut. Mortality over a 10-day period after surgery was less than 20% in both groups. The haemolymph of C. magister is slightly hyperosmotic to 100% sea water (Hunter & Rudy, 1975) and anuric crabs gained weight at a rate equal to the normal urinary rate in this medium (≃ 2% body weight/day, Holliday, 1978c). To prevent swelling, anuric crabs were weighed at 48 h intervals and were returned to their original body weights by removing haemolymph with a syringe. Serum and urine samples were taken over a 41-day period following surgery.

The effect of residence time of the urine in the bladder on urine magnesium concentration was measured using crabs with temporarily blocked nephropores. The urinary bladders of crabs in 100% sea water were drained and blocked with hot-melt glue as previously described (Holliday, 1978b). At intervals after this treatment the urinary seals were removed, a urine sample was taken and the seals were replaced.

Eyestalk ablation was performed as previously described (Holliday, 1978b) to determine the role of the eyestalk in controlling magnesium excretion. After ablation, crabs were returned to 100% sea water for 3 days, at which time serum and urine samples were taken. Soon after eyestalk ablation the hypodermis at the ischio-meral joints of the pereiopods changed colour from black to red, indicating that the x-organ sinus glands, which are known to exert at least partial hormonal control of chromato-phore function (Kleinholz, 1976) were removed. Mortality was less than 5%.

In vitro measurements of transepithelial electrical potential difference, resistance and short circuit current (TEP, R and SCC, respectively) were made in a lucite Ussing chamber. Crabs were immobilized by destruction of the thoracic ganglia, and the dorsal carapace and underlying hypodermis were cut away. Pieces of epigastric bladder tissue approximately 2 cm² in area were dissected as flat sheets under sea water, mounted between plastic coverslips with centred, bevelled-edge holes 0.80 cm² in area and clamped between the halves of the Ussing chamber. The chambers on either side of the tissue held 25 ml of CR which was vigorously stirred by a stream of air. Miniature calomel half-cells were positioned 1 cm from either side of the tissue in the cell and connected to a high impedance millivolt meter/voltage clamp device (Menninger, 1972). The tissue was set up in the apparatus and allowed to equilibrate for 10 min before measurements were begun. For resistance measurements, a current of 25/μA was applied across the tissue with Ag-AgCl electrodes and the resulting change in TEP was measured. Resistance was calculated using Ohm’s law and measurements were corrected for the resistance of the apparatus without the tissue in place.

Oxygen consumption by pieces of excised epigastric bladder tissue in CR was measured at intervals over a 3 h period at 12 °C in a Gilson DRP-14 respirometer using standard manometric techniques.

The unidirectional fluxes of ²⁸Mg, ²²Na and ³⁶Cl across matched pairs of excised bladder tissues were measured in a previously described, dual lucite chamber (Holliday, 1978 b). Pairs of epigastric bladder tissues from the same crab were excised under sea water and mounted between plastic coverslips with centred holes as described above. When clamped between the halves of the dual chamber, each tissue was supported between two separate chambers, each containing 1.00 ml of CR. The CR in each of the four chambers was aerated and stirred by a four-channel peristaltic pump at the rate of 2 ml/min. Isotope was added to the serosal-side chamber of one tissue and to the urine-side chamber of the other tissue to a final specific activity of 1000-3500 cpm/µequiv and the contents of the four chambers were sampled 1, 2 or 3 h later. Because the TEP was small and transient (Results), tissues were not short-circuited during flux measurements. Unidirectional isotope fluxes were calculated as the rate of appearance of the isotope in the two chambers which were originally isotope-free and are expressed as µequiv cm^-2 bladder h^-1. Preliminary experiments indicated that under these conditions the unidirectional ion fluxes were nearly stable for 6 h or more, and that the preparation was impermeable to the fluid volume marker, Glofil (® Abbott Laboratories, Chicago, Ill.). In one set of experiments, tissues were set up in the chamber and maintained under a nitrogen atmosphere for 1 h before and during ²²Na flux measurements.

Radioactivity was measured using standard solid (²⁸Mg, ²²Na) and liquid (³⁶Cl) scintillation techniques. Gamma emissions were counted with a Picker Nuclear 2840 E sodium iodide crystal well detector connected to a Picker Nuclear 628-057 pulse height analyser and 628-145 rate meter or with a Canberra Industries Bicron sodium iodide crystal detector and Omega-I multichannel analyser. Counts from ²⁸Mg were corrected for decay (half-life 21-3 h) by counting a ²⁸Mg standard with each sample. Beta emissions from ³⁶C1 were counted with a Beckman LS-150 liquid scintillation counter and counts were corrected for quenching by the external standard channels ratio method. Magnesium concentrations in serum, urine, and sea-water media were measured using the spectrophotometric thiazole yellow method of Sky-Peck (1964), as modified by Hunter & Rudy (1975). Absorbance was measured with a Zeiss PMQ-2 spectrophotometer. The standard deviation of the mean of five determinations of a 100 m-equiv/1 magnesium standard was ± 1.6 m-equiv/1. Sodium concentrations in serum and urine were measured by flame photometry using a Coleman 6/20 Junior II spectrophotometer and flame photometer. The standard deviation of the mean of five determinations of a 400 m-equiv/1 sodium standard was+ 2 m-equiv/1.

²²NaCl and Na³⁶Cl were obtained from ICN Pharmaceuticals, Irvine, Calif. ²⁸MgCl₂ was obtained from Brookhaven National Laboratories, Upton, N.Y. All chemicals used in this study were of reagent grade.

The significance of the difference between mean values of experimental parameters was assessed using a paired or unpaired i-test, as appropriate. Probabilities of 5% or less were considered significant, those of 1 % or less were considered highly significant.

The values of serum magnesium (Mg_s) and urine magnesium (Mg_u) in crabs acclimated for 48 h in 100% and 75% sea water, together with values for magnesium-depleted crabs maintained in magnesium-free 75 % sea water for 6 days, are shown in Table 1. In 100% or 75% seawater, Mg_s is maintained at a level 60-75% less than that of the medium. Mg_u is much greater than Mg_s and the urine:serum ratio (U/S) is highest in 100% sea water, the medium in which the crab hyporegulates magnesium to the greatest extent. Magnesium secretion into the urine is under physiological control ; crabs kept in magnesium-free 75 % sea water became depleted of magnesium and stopped concentrating it in the urine (Table 1).

If, as the data shown in Table 1 would seem to indicate, the antennal gland is responsible for hyporegulation of Mg_a, then anuria should eliminate this hyporegulatcory ability. Mg_s in surgically anuric crabs maintained in 100% sea water rose steadily over a period of 26 days to a level 20-25 m-equiv/1 below that of the medium (Fig. 1), indicating that the antennal gland is responsible for most of the crab’s hyporegulatory ability. Sham-operated crabs showed a transitory rise in Mg_g 5 days after surgery but recovered full hyporegulatory ability.

Although the above experiments indicate that the antennal gland is responsible for magnesium hyporegulation, the antennal gland participates only slightly in elimination of injected magnesium loads (Table 2). Mg_s returned to control levels 24-36 h after magnesium loading (data not shown), but only 26% of the load was excreted in the urine. Urinary rate was not significantly changed by magnesium loading. Single, injected magnesium loads are apparently dealt with by extrarenal mechanisms.

Evidence that the bladder of the antennal gland is the site of magnesium secretion into the urine is provided by the data shown in Fig. 2. In crabs with initially empty bladders and blocked nephropores, Mg_u was a function of residence time of the urine in the bladder, indicating that the bladder, and not a more proximal part of the antennal gland, is responsible for secretion of magnesium into the urine. As Mg_u rose, urine sodium levels (Na_u) fell and there was a highly significant negative correlation between these variables (Fig. 3). Thus, net magnesium and sodium movements across the bladder epithelium may be linked in some manner. The ratio of the apparent exchange of magnesium for sodium across the bladder wall was calculated from the slope of the regression line and was approximately 1.5 equiv Mg: 1.0 equiv Na.

Eyestalk ablation had no significant effect on Mg_g and Mg_u (Table 3). Thus, it seems unlikely that magnesium excretion by the antennal gland is under hormonal control by a neurosecretory release site in the eyestalk.

The results of in vitro measurements of TEP, R and SCC in excised bladder tissue are shown in Table 4. TEP was small and lumen-negative when measured 10 min after the tissue was set up in the Ussing chamber, but the values rose to zero within i h. Preliminary experiments in which transepithelial fluxes of isotopes of magnesium, sodium and chloride were measured at hourly intervals for 6 h showed nearly stable values (data not shown). Thus, the small, transient TEP is apparently not associated with the movements of these ions across the tissue. Values of R and SCC were also small. Excised bladder tissue is metabolically active, as oxygen consumption by the tissue was relatively high (Table 4) and was stable over the 3 h period of measurement. Dehnel & McCaughran (1964) and Hulbert, Schneider & Moon (1976) have obtained similar values for oxygen consumption by crab gills.

When measured in the dual lucite cell, unidirectional fluxes of ²⁸Mg across matched pairs of bladder tissue were relatively low, but there was a highly significant, net serosa-to-lumen (SL) flux of the isotope (Table 5). Crabs maintained in magnesium-free 75% sea water stopped concentrating magnesium in the urine (Table i) and the SL flux of “Mg was greatly reduced, with the result that the difference between the unidirectional fluxes was not significant. Thus, the bladder is a site of magnesium secretion into the urine. Ouabain (5 mm, on both sides of the tissue), a specific inhibitor of Na, K-ATPase-powered sodium transport (Dahl & Hokin, 1974), had no significant effect on the lumen-to-serosa (LS) ²⁸Mg flux but reduced the SL flux and eliminated the significant difference between the unidirectional fluxes. This suggests that magnesium transport by the bladder is linked in some manner to sodium transport.

Unidirectional fluxes of ²²Na acros excised bladder tissue were much greater than ²⁸Mg fluxes (Table 5) and there was a significant, net LS flux. Surprisingly, neither unidirectional nor net ²²Na fluxes were significantly reduced by 5 mm ouabain on both sides of the tissue. The significant net LS ²²Na flux was, however, abolished by a large decrease in the LS flux when flux measurements were made under a nitrogen atmosphere. Thus, although the net ²²Na flux is apparently not directly associated with a Na, K-ATPase mechanism it is dependent on aerobic metabolism.

Unidirectional fluxes of ³⁶Cl across excised bladder tissue were much greater than the corresponding ²²Na fluxes (Table 5). However, the significant net LS ³⁶Cl flux was nearly equal to the net ²²Na flux and was in the same direction.

Most marine crustaceans hyporegulate Mg_s and, because U/S ratios for magnesium are usually much greater than inulin U/S ratios, the antennal gland is thought to be responsible for this hyporegulatory ability (Riegel & Cook, 1975). This hypothesis is confirmed in the present study; Mg_s in surgically anuric crabs slowly rose to a level 20-25 m-equiv/1 below that of the medium (Fig. 1), indicating that the antennal gland is responsible for most of the crab’s capacity for magnesium hyporegulation. It could be argued that non-permeant haemolymph proteins or an inside-positive TEP could account for the remaining small difference between Mg_s and the magnesium level in the external medium. However, Webb (1940) found that dialysis of haemolymph from the crab, Carcinus maenas, against sea water resulted in slightly higher levels of magnesium in the dialysate than in sea water. Further, Hunter (1973) has found that C. magister has a small, inside-negative TEP in 100% sea water. Thus, the observation that Mg_s remained below that of the seawater medium in anuric crabs may indicate that the crab has some ability to secrete magnesium extrarenally.

There is extensive evidence that Mg_u in crustaceans is under physiological control. Crustaceans commonly stop concentrating magnesium in the urine when Mg_s is lowered by acclimation to dilute or magnesium-free media (Gifford, 1962; Dehnel & Carefoot, 1965; Lockwood & Riegel, 1969; Hunter & Rudy, 1975; Franklin et al. 1978), and C. magister is typical in this respect (Table 1). Further, several crustaceans apparently reabsorb magnesium from the urine under these conditions (Scholles, 1933; Burger, 1957; Gifford, 1962; DeLeersnyder & Hoestelandt, 1963; De Leersnyder, 1967; an unnamed crustacean mentioned by Riegel, 1972, p. 121). Conversely, several crustaceans respond to acclimation to high-salinity or magnesium-enriched media by elevating Mg_u (Webb, 1940; Prosser, Green & Chow, 1955; Green et al. 1959; Gross & Marshall, 1960; Dehnel & Carefoot, 1965 ; Gross & Capen, 1966; Gross et al. 1966), and a few studies have demonstrated that the increase was not caused by increased fluid reabsorption from the urine (Lockwood & Riegel, 1969; Franklin et al. 1978). Thus, magnesium secretion into the urine is under physiological control and may be increased, curtailed, or reversed as necessary to regulate Mg_s. In the present study, single, injected magnesium loads did not greatly increase Mg_u, and only 26% of the injected load was excreted in the urine (Table 2). Several authors have reported that injected magnesium loads are not excreted by the antennal glands (Bialaszewicz, 1931; Burger, 1957; Dall, 1967). The crab, C. maenas, has been reported to eliminate injected magnesium loads via the antennal glands (Lockwood & Riegel, 1969), but these authors did not calculate the fraction of the load excreted in the urine. Thus, C. meanas may be similar to C. magister in this respect. In any case, short-term elevation of Mg_s was not a sufficient stimulus to greatly increase urinary magnesium excretion in C. magister; it may be that a longer period of elevated Mg_s is necessary for this to occur.

The bladder has been suggested as the site of magnesium secretion into the urine; Gross & Capen (1966) have shown that Mg_u is a function of the residence time of the urine in the bladder in Pachygrapsus crassipes. This is also the case in C. magister (Fig. 2). As noted by Gross & Capen (1966), if magnesium were secreted into the urine by a more proximal part of the antennal gland, then urine entering the bladder would already have a high magnesium concentration and Mg_u would not increase with residence time of the urine in the bladder.

The observed ratio of apparent opposing in vivo movements of magnesium and sodium across the bladder of 1.5 equiv Mg:ro equiv Na (Fig. 3) might be thought to be electrogenic. However, as noted by Hunter & Rudy (1975), magnesium salts do not ionize as completely as sodium salts. When the appropriate activity coefficients of MgSO₄ and MgCl_a (0.2-0.5) and NaCl (0.7, all figures from Hamed & Owen, 1958) at the observed urinary concentrations are applied, the ratio is reduced to approximately unity. Further, excised bladder tissue showed only a small, transient TEP (Table 4) and Franklin et al. (1978) have shown that only a small TEP across the bladder is present in Palaemon serratus. Thus, ion transport by crustacean bladder is apparently not, or only slightly, electrogenic, a characteristic which this tissue shares with other ‘leaky’ epithelia such as vertebrate gallbladder (Rose, 1978). On the basis of in vivo experiments a direct (1 mol Mg: 1 mol Na) exchange in the antennal gland has been both rejected (Riegel & Lockwood, 1961) and calculated to be possible (Gross & Capen, 1966; Hunter & Rudy, 1975). However, both fluid reabsorption from the urine and magnesium secretion into it occur in marine crustaceans. Further, there is good evidence that these two processes affecting the final level of magnesium in the urine are separate (Franklin et al. 1978). Thus, it seems likely that the apparent Mg: Na exchange in the antennal gland is less the result of a specific exchange transport mechanism than the net result of at least two different processes (Hunter, 1973; Franklin et al. 1978).

The present study is the first in which magnesium transport by the crustacean bladder has been directly demonstrated; bladder tissue shows a highly significant, net SL flux of 0.280 µequiv Mg cm^-2 h^-1 (Table 5). The surface area of the bladder in C. magister is approximately 0.9 cm²/g body weight (Holliday, 1978b). Thus a 500 g crab (with ≃ 450 cm² of bladder) could transport ≃ 3 m-equiv of magnesium/ day into the urine for excretion. This calculation is simplistic in that it does not take into account the effects of factors such as imposition of a normal LS magnesium gradient across the bladder or lack of metabolic substrates on the in vitro rate of net SL magnesium transport. However, the calculated value is in reasonable agreement with the measured in vivo value of 1.80 m-equiv magnesium excreted/day (Table 2, mean weight of crabs was 475 g) ; thus the observed rate of magnesium transport into the urine by the bladder in vitro is sufficient to account for magnesium excretion by the entire antennal gland in vivo. Excised bladder tissue from magnesium-depleted crabs showed greatly reduced SL and net fluxes of magnesium (Table 5), a finding in accord with the observed cessation of magnesium secretion into the urine in such crabs in vivo (Table 1).

In vitro net magnesium fluxes are greatly inhibited by 5 mm ouabain on both sides of the tissue (Table 5). Preliminary, unpublished experiments indicate that homogenates of bladder tissue from C. magister have moderate levels of Na, K-ATPase activity ( ≃ 10 µmol P₁ mg^-1 protein h^-1 at 37 °C). Thus, magnesium transport by the bladder may be linked to sodium transport, and it is possible that magnesium enters the bladder cells by a sodium-coupled process at the serosal side of the bladder and is extruded by a magnesium transport mechanism at the luminal side. This hypothesis is supported by the unpublished observation that magnesium levels in bladder tissue are about the same as Mg_s (≃ 30 m-equiv/1 tissue water as opposed to 39 m-equiv/1 serum).

The finding in the present study that magnesium transport into the urine in vitro is inhibited by ouabain would seem to be at variance with the data of Gross & Capen (1966, p. 282). These authors introduced perfusion fluid (of roughly equivalent ionic composition to the CR used in the present study) containing 20 m-equiv Mg/1 into the drained, rinsed bladders of P. crassipes and sampled the fluid after 24 h. When ouabain (0.5 or 1 mm) was present in the perfusion fluid, magnesium was still transported into the perfusion fluid, as the magnesium concentration rose from 20 m-equiv/1 to 240 (s.D. 178) m-equiv/1. This finding is curious, as the magnesium concentration in perfusion fluid without ouabain which had been in the bladder 24-48 h in an apparently identical experiment was only 129 (s.D. 55) m-equiv/1 (Gross & Capen, 1966, p. 279), indicating that ouabain nearly doubled the rate of magnesium transport into the perfusion fluid in the bladder, although the mean values in the two experiments are not significantly different. The lack of significant ouabain inhibition of magnesium secretion with luminal exposure in the bladder of P. crassipes is consistent with the sodium-coupled mechanism proposed above for the bladder of C. magister; only serosal exposure to ouabain would be expected to inhibit the Na, K-ATPase-powered sodium pump which has a basolateral location in nearly all animal tissues examined to date (DiBona & Mills, 1979). Gross & Capen (1966) did note that ouabain had apparently diffused out of the bladder into the crabs’ haemolymph and that ouabain was, therefore, present on both sides of the bladder. However, the haemolymph volume of crabs is much larger than the bladder volume (Riegel et al. 1974) and, thus, the concentration of ouabain in the haemolymph was probably too low to significantly inhibit magnesium transport by the bladder.

The large, net LS sodium flux across excised bladder tissue was dependent on aerobic metabolism and is presumably an active process, but 5 mm ouabain on both sides of the tissue had no significant effect on sodium fluxes (Table 5). There is good evidence that magnesium secretion and fluid reabsorption in the antennal gland are separate processes (Franklin et al. 1978). It is tempting to speculate that only a small fraction of the large, net LS sodium flux (Table 5) across the bladder in vitro is associated, through maintenance of electrical neutrality across the tissue, with the opposing, ouabain-sensitive movement of magnesium. Using the ratio of the increase in Mg_u to the decrease in Na_u of 1.5 equiv Mg: 1.0 equiv Na found in the in vivo experiment with blocked nephropores (Fig. 3), one would expect a decrease of only 0.19 µequiv Na cm^-2 h^-1 in the net ²²Na flux when ouabain is present. This decrease is too small to cause a significant decrease in the large, net LS ²²Na flux (Table 5). The large, remaining, net LS sodium flux, which is ouabain-insensitive, or the nearly equal, net LS flux of chloride, may be the driving force for fluid reabsorption from the urine by the antennal gland. This hypothesis would account for the lack of a significant effect of ouabain on the net LS ²²Na flux and still accommodate the ouabain-sensitivity of the net SL flux of ²⁸Mg. If this is the case, then fluid reabsorption by crab bladder is fundamentally different from the process in vertebrate tissues such as gallbladder, where fluid reabsorption is ouabain-sensitive (Frizzell, Field & Schultz, 1979). Mykles (1980) has shown that fluid absorption from the lumen of the midgut in the lobster, Homarus americanus, is insensitive to 0.1 mm ouabain, requires the presence of sodium in the medium and is abolished by 1 mm cyanide. These findings are consistent with the hypothesis that fluid reabsorption by the antennal gland could be driven by the ouabain-insensitive, anoxia-inhibited, net LS sodium flux observed in vitro in the bladder of C. magister (Table 5).

Little work has been done on the mechanism of control of Mg_u in crustaceans. Franklin et al. (1978) have presented preliminary evidence that magnesium secretion into the urine in the prawn, Palaemon serratus, is under hormonal control by the eyestalk. This is apparently not the case in C. magister, as eyestalk ablation had no significant effect on Mg_s or Mg_u (Table 3). Although there is excellent evidence for the existence of a hyperglycaemic hormone in the crustacean eyestalk, the hormone is different in macrurans and brachyurans (Kleinholz & Keller, 1973) and the latter often do not respond to deficiency (i.e. eyestalk ablation) experiments as would be expected (Kleinholz, 1976; Holliday, 1978b). It may be that there is a fundamental difference in the nature of control of magnesium secretion in these two groups. Obviously, much work remains to be done in this area.

The data presented in the present study confirm the hypothesis of Gross & Capen (1966) that magnesium is secreted into the urine by the bladder in marine crabs and are consistent with the hypothesis that fluid reabsorption also occurs in the bladder. Resolution of the role, if any, of the other parts of the antennal gland in these processes awaits the development and application of suitable micropuncture techniques.

This work represents part of a Ph.D. dissertation submitted to the Graduate School of the University of Oregon. The author wishes to express thanks to Dr Paul P. Rudy for his help and guidance throughout the course of the research and Drs David S. Miller and Bodil Schmidt-Nielsen for stimulating discussions and comments on the original manuscript. This work was supported in part by B.R.S.G. Grant RR07080 awarded to the Graduate School of the University of Oregon by the Biomedical Research Support Grant Program, Division of Research Resources, National Institutes of Health.