Feeding and digestion in low salinity in an osmoconforming crab, Cancer gracilis I. Cardiovascular and respiratory responses

The osmoregulatory physiology of decapod crustaceans has been studied extensively (for reviews, see Mantel and Farmer, 1983; Pequeux,1995; Charmantier et al.,2001). Within this field there is a growing body of literature on the cardiovascular and respiratory responses of decapod crustaceans to low salinity. The most common response to low salinity is a pronounced tachycardia(Hume and Berlind, 1976; Cumberlidge and Uglow, 1977; Spaargaren, 1982; McGaw and McMahon, 1996; McGaw and McMahon, 2003; McGaw and Reiber, 1998; Dufort et al., 2001; McGaw, 2006a). This is coupled with an increase in oxygen uptake (Dehnel,1960; King, 1965; Engel et al., 1975; Taylor, 1977; Guerin and Stickle, 1992; Jury et al., 1994), which is driven by an increased scaphognathite beat frequency(McGaw and McMahon, 1996; Dufort et al., 2001). These changes are thought to reflect an increased energy requirement for active ion uptake (Taylor, 1977; Jury et al., 1994) as well as increased activity levels of crabs in low salinity(McGaw et al., 1999). Most of these studies have concentrated on decapod species that are classified as either efficient or weak hyperosmoregulators. Studies on cardiac and respiratory responses of osmoconforming species of decapod crustaceans are very limited. In contrast to species that are able to hold the body fluid concentration above that of the medium, osmoconforming species exhibit a bradycardia (Spaargaren, 1973; Cornell, 1973; Cornell, 1974; Cornell, 1979) and a decrease in oxygen uptake (Savant and Amte,1995).

All the crustaceans used in these previous experiments were starved before and/or were not fed during experiments. This protocol is adopted since the stimulatory effect of food ingestion (specific dynamic action) on metabolic processes is well known (Wang,2001). Decapod crustaceans are no exception; oxygen uptake increases immediately after feeding and reaches maximal levels within 1 to 4 h(Houlihan et al., 1990; McGaw and Reiber, 2000; Robertson et al., 2002; Mente et al., 2003). Oxygen uptake can remain elevated for over 48 h(Legeay and Massabuau, 1999; McGaw and Reiber, 2000). Blood flow is diverted to the muscles while feeding and to the digestive organs thereafter (McGaw and Reiber,2000; McGaw, 2005; McGaw, 2006a).

The digestive state of an organism can be very important. Recent work on osmoregulating crustacean species (Carcinus maenas, Cancer magister)shows that digestion can pose additional demands on physiological systems,leading to an increased mortality rate of postprandial crabs in low salinity(Legeay and Massabuau, 2000; McGaw, 2006a). Because crustaceans are not normally starved before encountering low salinity (D. L. Curtis and I. J. M., unpublished observation) the question then arises as to how an animal balances the simultaneous demands of these physiological systems.

The graceful crab Cancer gracilis inhabits sandy and muddy bays along the Pacific coast of North America. Cancer gracilis is an osmoconformer and cannot survive in salinities below 55%SW, although it can be exposed to reduced salinities from freshwater runoff for several hours at a time (D. L. Curtis, E. K. Jensen and I. J. McGaw, submitted for publication). Cancer gracilis becomes quiescent in low salinity and attempts to isolate the branchial chambers from the surrounding water. This results in a decrease on oxygen uptake while the branchial chambers are sealed (D. L. Curtis, E. K. Jensen and I. J. McGaw, submitted for publication). Digestive processes elicit the opposite response(McGaw and Reiber, 2000; Robertson et al., 2002) and thus may pose an additional burden to crabs already attempting to regulate oxygen uptake and blood supply to the tissues(McGaw, 2005; McGaw, 2006a). It was hypothesized that because osmoconforming species have limited ability to cope with hyposaline exposure they would display different reactions to the simultaneous demands of digestion and low salinity compared with osmoregulating species, such as Cancer magister(McGaw, 2006a). Both respiratory and cardiac responses were measured: oxygen uptake was used as a basic measure of metabolic rate, while changes in heart rate provided an important measure of stress in crustaceans(Handy and Depledge, 1999). However, heart rate alone is not an accurate means of assessing of the total amount of haemolymph (cardiac output) delivered to the system(McGaw and McMahon, 1996). The cardiac output is also dependent on stroke volume of the heart, which can vary independently of heart rate (McGaw and McMahon, 1996). The haemolymph pumped from the heart is directed through a complex series of arteries and capillary-like vessels and decapod crustaceans are able to regulate the amount delivered through each arterial system (McGaw and Reiber,2002). This is not just important for efficient delivery of nutrients and gases; the diversion of flow to metabolically active tissues may enhance the ability of crabs to cope with exposure to low salinity(McGaw and McMahon, 1996; McGaw and McMahon, 2003; McGaw and Reiber, 1998). Therefore, the aim of the present study was twofold: (1) to investigate the respiratory and cardiac responses of an osmoconforming species of decapod crustacean to hyposaline exposure, and (2) to determine how digestive processes affect the ability of Cancer gracilis to balance the demands of physiological systems in low salinity environments and whether they prioritize or exhibit additivity (mix) of physiological responses(Bennett and Hicks, 2001; Hicks and Bennett, 2004).

Adult male, intermoult graceful crabs Cancer gracilis Dana of 225-280 g were trapped in the Bamfield Inlet, British Columbia, Canada,between May and August 2005. They were transferred to the Bamfield Marine Sciences Centre and held in running seawater (SW; 31-32‰ at 11±1°C) for a week prior to experimentation. Crabs were fed fish every other day, but were isolated from the general population and starved for 3 days prior to experimentation. This time period allowed all food to be evacuated from the digestive system, but avoided large-scale physiological changes associated with starvation(Wallace, 1973).

A 545C pulsed-Doppler flowmeter (University of Iowa-Bioengineering, Iowa City, IA, USA) was used to measure haemolymph flow rates in each of the major arterial systems exiting from the heart. Piezo-electric Doppler flow probes were either implanted directly above the arteries in grooves abraded in the carapace (anterior and posterior aortae, anterolateral arteries) or guided to lie adjacent to the artery via internal catheter-mounted probes(hepatic arteries, sternal artery) and held in place with dental wax and super glue. Maximal signal was obtained by using the fine depth focus on the Doppler machine. Following experimentation the animals were sacrificed for measurement of arterial diameters and verification of probe implants. Heart rate was obtained by counting the peaks on the phasic flow traces; summation of all arterial flows (paired arteries were doubled) gave a value for cardiac output and division of this value by heart rate yielded cardiac stroke volume. A detailed description of the set-up and methods is covered elsewhere(Airriess et al., 1994). Changes in pressure were measured in the branchial chambers allowing calculation of scaphognathite beat frequency (ventilation rate). Holes drilled into the carapace above the branchial chamber were covered with dental dam and a chronically implanted polythene catheter (PE 160) was held in place with wax and super glue. The catheter was filled with seawater and connected to a disposable blood pressure transducer (MLT0698, ADInstruments, Mountain View,CA, USA). During experiments, crabs were held in a circular tank of 320 mm diameter×300 mm depth in aerated seawater at 10-12°C, and a layer of gravel lined the bottom. The tips of the claws were glued together to prevent the crabs from cutting the catheters, but other than this they were able to move freely. Data for cardiac and ventilatory parameters were recorded continuously using an ADInstruments data acquisition system.

Oxygen uptake was measured using a Qubit D101 intermittent flow respirometry system (Kingston, ON, Canada). Crabs (N=10) were held in a cylindrical chamber of 200 mm diameter×80 mm depth. Oxygen uptake was calculated at 30 min intervals during a 10 min decline in oxygen levels while the chamber was sealed, then the chamber was continuously flushed between readings. Oxygen uptake was recorded on a Loligo data acquisition system(Copenhagen, Denmark).

The crabs were allowed to settle for 12 h in the chambers before experimentation. All recordings were carried out in constant dim light, which helped reduce any nocturnal activity. During experiments the apparatus was surrounded by black plastic sheeting to avoid visual disturbance to the animal. The salinity was changed by draining part of the tank (without aerially exposing the crab), and adding a known volume of freshwater at ambient temperature and oxygen levels. Salinity was checked using a YSI 30 conductivity meter (Yellow Springs, OH, USA); for reference 100%SW=32‰salinity. New steady states of salinity were reached in the experimental apparatus within 10 min and did not vary by more than 0.1‰ during experiments. A salinity regime of 65%SW (approximately 21‰) was used since this level was above what is considered as a survivable salinity (55%SW)for Cancer gracilis, but also within the range that this species adopts an isolation response (D. L. Curtis, E. K. Jensen and I. J. McGaw,submitted for publication). The time course of low salinity exposure was chosen to emulate naturally occurring conditions based on the tidal cycle in Barkley Sound, British Columbia. For feeding, a polythene tube (PE160) was inserted into the oesophagus. This allowed a liquified meal of fish muscle,equal to 2% of the crab's body mass, to be administered at a rate of approximately 2 ml min^-1. It also allowed feeding time to be synchronized.

Four separate experiments were carried out. In the first experimental series, cardiovascular and respiratory parameters were monitored for a 3 h control period in 100%SW (32‰). The crabs (N=10) were then fed and changes monitored for a further 12 h in 100%SW. In the second experimental series starved crabs (for 3 days) were monitored for 3 h in 100%SW, the salinity was then lowered to 65%SW and cardiac and respiratory parameters measured for 6 h, after which 100%SW was restored for a further 6 h. In the third set of experiments, cardiovascular and respiratory parameters of crabs(N=10) were monitored for 3 h in 100%SW. The animals were then fed; 3 h after feeding, low salinity (65%SW) was initiated for a total time of 6 h. Full-strength seawater was then restored for a further 6 h. In a final series of experiments, ten crabs that had been fed 21 h previously were monitored in control conditions for 3 h. Low salinity was then initiated for 6 h, after which 100%SW was restored for an additional 6 h. Values are presented as the mean ± s.e.m. (standard error of the mean).

One-way ANOVA with repeated measures design was used to test for significant differences in cardiovascular and ventilatory parameters. Data showing a significant effect, were further analyzed by a Fisher's LSD multiple comparison test (P<0.01) to determine at which time periods significant effects were observed.

Cancer gracilis maintained its heart rate between mean levels of 71 and 75±2.1 beats min^-1 during control conditions(Fig. 1A). Heart rate increased significantly during feeding, reaching 85.6±2.4 beats min^-1(ANOVA, F=2.86, P<0.001); there was a slight decrease immediately after feeding, but heart rate remained elevated for 2 h after feeding before returning to control levels. There was a twofold increase in stroke volume during feeding (ANOVA, F=3.49, P<0.001),reaching 0.298±0.046 ml beat^-1. This was followed by an immediate decrease in stroke volume back to pre-feeding values(Fig. 1B). A combined increase in heart rate and stroke volume lead to an increased cardiac output during feeding (ANOVA, F=5.59, P<0.001). There was a decrease in cardiac output after feeding but it remained elevated over pre-feeding levels for 5 h with a further slight increase at 14 h(Fig. 1C).

Haemolymph flow rates through the arterial systems were somewhat variable(Fig. 1D-H). Despite apparent trends, the only statistically significant change in flow rate occurred through the sternal artery (ANOVA, F=5.55, P<0.001). Haemolymph flow rates in the sternal artery increased from mean pre-feeding levels of between 6 and 6.8 ml min^-1 up to 18.6±3.9 ml min^-1 during feeding (Fig. 1H). Thereafter, flow rates returned to pre-treatment levels;however, at 6 h there was a slight but statistically significant increase to 10.2±2.1 ml min^-1 and levels remained elevated over prefeeding levels until 10 h, with an additional significant increase at 14 h.

Ventilation rate increased significantly from pre-feeding levels of 81-83 beats min^-1 to 93±3.3 beats min^-1 during feeding(Fig. 2A; ANOVA, F=2.12, P=0.01). The ventilation rate remained significantly elevated for 3 h after feeding before dropping back to pre-treatment levels. Oxygen uptake was maintained between 28-35 mg O₂ kg^-1h^-1 during pre-feeding conditions(Fig. 2B). The oxygen uptake doubled during feeding, reaching 71.2±8.2 mg O₂kg^-1 h^-1 (ANOVA, F=7.44, P<0.001). There was a slight decline thereafter, but oxygen uptake was still elevated over pre-feeding levels, 48 h after feeding. Pre-treatment levels were regained between 55-60 h after feeding (not shown).

Exposure to 65%SW resulted in a pronounced bradycardia in Cancer gracilis. Heart rate decreased from mean levels of 72-74 beats min^-1 to between 45 and 50 beats min^-1(Fig. 3A; ANOVA, F=22.77, P<0.001). Pre-treatment levels were regained within 1 h of return to 100%SW. Stroke volume increased significantly during the first hour of low salinity exposure from mean levels of 0.25-0.26 ml beat^-1 to 0.34±0.07 ml beat^-1(Fig. 3B; ANOVA, F=2.55, P=0.002), after which it dropped back to pre-treatment levels. Upon return to 100%SW there was a further decrease below pre-treatment levels, which was sustained for the 6 h recovery period. Cardiac output was maintained between mean levels of 18 and 19 ml min^-1during pre-treatment conditions (Fig. 3C). There was a sustained decrease in cardiac output in low salinity (ANOVA, F=15.13, P<0.001) down to 10.6 ml min^-1. Cardiac output recovered slowly upon return to full-strength seawater; recovery of pre-treatment levels started after 2 h.

Haemolymph flow rates through the anterior aorta decreased slightly after 3-4 h exposure to low salinity (Fig. 3D; ANOVA, F=2.88, P<0.001). When the crabs were returned to 100%SW a significant increase in flow through the anterior aorta occurred (over both pre-treatment and treatment conditions) and was sustained for 2 h. There was also a significant decrease in flow rates through the anterolateral arteries during hyposaline exposure(Fig. 3E; ANOVA, F=2.78, P<0.001): haemolymph flows dropped from mean levels of 0.68-0.75 ml min^-1 to 0.56±0.1 ml min^-1. These levels were sustained for the period of exposure to low salinity. Recovery of pretreatment levels occurred within 2 h of return to 100%SW. Although there was an apparent decrease in flow rates through the hepatic arteries (Fig. 3F), no statistical significance could be demonstrated (ANOVA, F=0.92, P>0.05). Flow rates through the posterior aorta were maintained between 0.32 and 0.38 ml min^-1 in control conditions(Fig. 3G). After 1 h in 65%SW there was a significant decrease, and flow rates reached 0.24 ml min^-1. This decrease was sustained for the low salinity period(ANOVA, F=2.87, P<0.001). Pre-treatment levels were regained within 2 h of return to 100%SW. The most pronounced changes were observed in the sternal artery (Fig. 3H). Haemolymph flow rates decreased significantly from between 11.3 and 12.2 ml min^-1 to between 5.1 and 6.1 ml min^-1in low salinity (ANOVA, F=17.87, P<0.001). 2 h after return to 100%SW haemolymph flows had increased significantly over low salinity treatment values. However, these levels were still significantly lower than those measured during pre-treatment conditions.

Ventilation rates in pre-treatment conditions were maintained between 75.5 and 78 beats min^-1. These dropped significantly to 61-65 beats min^-1 in low salinity (ANOVA, F=8.11, P<0.001)and were maintained for the duration of 65%SW exposure. Pre-treatment levels were regained after 2 h in 100%SW (Fig. 4A). In control conditions oxygen uptake was maintained between 38 and 42 mg O₂ kg^-1 h^-1(Fig. 4B). There was a sharp decrease during low salinity exposure, reaching 8.2±2.4 mg O₂ kg^-1 h^-1 (ANOVA, F=9.01, P<0.001). Oxygen uptake increased steadily thereafter and pre-treatment levels were regained at 6 h. Upon return to 100%SW there was a rapid increase in oxygen uptake reaching over 70 mg O₂kg^-1 h^-1. Oxygen uptake remained elevated over pre-treatment levels for 5 h before dropping back to pre-treatment levels.

Cancer gracilis was exposed to low salinity 3 h after feeding while food was still in the foregut and at 24 h after food intake when the gut was cleared and intracellular digestion was underway(McGaw, 2006b). In control conditions Cancer gracilis maintained its heart rate between 70.2 and 71.8 beats min^-1 (Fig. 5A). There was a trend towards an increase in heart rate with feeding, but this proved to be statistically insignificant (Fisher's LSD P>0.01). A significant drop in heart rate down to 64.9±3.7 beats min^-1 occurred during the first hour of low salinity exposure(ANOVA, F=7.66, P<0.001). Thereafter, heart rate increased slowly and at 10 h it was significantly higher than that measured during pre-treatment conditions. Upon return to 100%SW there was a further increase; within 2 h heart rate reached 87±2.8 beats min^-1and it remained elevated over both pretreatment and treatment conditions for the experimental recovery period. The changes in stroke volume were less pronounced. There was a slight, but steady increase in stroke volume following feeding; at 6 h it was significantly higher than pre-feeding levels. A further transient increase in stroke volume occurred during the first hour of low salinity exposure, reaching 0.27±0.05 ml beat^-1 (ANOVA, F=2.3, P<0.01) followed by a subsequent decline back to pretreatment values. After 2 h recovery in 100%SW, stroke volume increased over levels measured during pre-feeding and low salinity exposure. In control conditions cardiac output was maintained between 12.4 and 14.1 ml min^-1 (Fig. 5C). Cardiac output started to increase after feeding and it was significantly elevated over pretreatment levels at 6 h (ANOVA, F=5.12, P<0.001). There was no significant change in cardiac output during low salinity exposure. During the recovery phase in 100%SW cardiac output increased to over 20 ml min^-1. This was significantly higher than levels recorded during pre-feeding and low salinity treatments.

Haemolymph flow rates also changed in some of the arterial systems. There was no significant change in haemolymph flow rates through the anterior aorta(Fig. 5D; ANOVA, F=1.69, P>0.01). There was also no significant change in flows through the anterolateral arteries during feeding or low salinity treatment (Fig. 5E); flow rates were maintained between 1 and 1.4 ml min^-1. However, within 2 h of return to 100%SW anterolateral artery flows increased to over 2 ml min^-1 (ANOVA, F=9.47, P<0.001). These were significantly higher than those measured during pre-feeding and treatment conditions. A similar pattern was observed in the posterior aorta(Fig. 5G). Food ingestion and low salinity had no effect, but a significant increase in flows occurred during the recovery period in 100%SW (ANOVA, F=3.32, P<0.001). Haemolymph flow rates through the hepatic arteries were more variable (Fig. 5F). A slight but significant decrease in flows occurred after 4 h in low salinity(ANOVA, F=2.16, P<0.01). During the recovery phase in 100%SW flows were significantly elevated above pre-treatment and treatment conditions. Haemolymph flows through the sternal artery varied between 6.7 and 7.3 ml min^-1 in control conditions. Flow rates had risen significantly by the end of the feeding period (ANOVA, F=2.43, P<0.01) and remained elevated during low salinity exposure. There was a further increase in flows over pretreatment levels, 2 h after return to 100%SW.

Ventilation rate was maintained between 74.8 and 76.9 beats min^-1 in control conditions(Fig. 6A), and increased significantly 2 h after feeding (ANOVA, F=2.34, P<0.01). No significant changes were observed during exposure to low salinity. After 3 h recovery in 100%SW ventilation rate reached 84 beats min^-1,significantly higher than pre-treatment levels (Fisher's LSD, P<0.01). Oxygen uptake was maintained between 23.5 and 29.8 mg O₂ kg^-1 h^-1 in control conditions(Fig. 6B). There was a threefold increase in oxygen uptake during feeding, followed by a subsequent decrease to approximately 50 mg O₂ kg^-1 h^-1(ANOVA, F=20.41, P<0.001). Upon treatment with 65%SW oxygen uptake decreased to 12.2±3.5 mg O₂ kg^-1h^-1; feeding levels were regained within 1.5 h. There was an increase during the recovery period in 100%SW and oxygen uptake increased over feeding and salinity treatment values, reaching 82.1±8.5 mg O₂ kg^-1 h^-1.

A similar pattern in cardiovascular and respiratory parameters occurred when low salinity was administered 24 h after feeding. In control conditions heart rate was maintained between mean levels of 71 and 72.6 beats min^-1 (Fig. 7A). A transient decrease to 62.9 beats min^-1 occurred during the first hour of low salinity exposure (ANOVA, F=3.7, P<0.001). Pre-treatment levels were quickly regained and maintained for the low salinity treatment period. After 3 h recovery in seawater heart rate had risen over both pre-treatment and low salinity treatments, reaching maximal values of 85±3.4 beats min^-1. The stroke volume was maintained between 0.22 and 0.26 ml beat^-1 in control conditions(Fig. 7B). There was no change in stroke volume during the first 3 h of low salinity exposure, but by 4 h stroke volume had decreased below pre-treatment levels, reaching 0.15±0.017 ml beat^-1 (ANOVA, F=3.88, P<0.001). Pretreatment levels were regained after 2 h in 100%SW. Cardiac output was maintained between 16 ml min^-1 and 18.5 ml min^-1 in pre-treatment conditions(Fig. 7C). After 4 h exposure to 65%SW, cardiac output had decreased to 13.1±1.7 beats min^-1 (ANOVA, F=4.09, P<0.001). Pretreatment levels were regained after 3 h recovery in 100%SW.

Haemolymph flow rates through the anterior aorta were variable and no significant change during the salinity treatment or the recovery phase could be demonstrated (Fig. 7D;ANOVA, F=1.05, P>0.05). There was a short-term significant decrease in flow rates through the anterolateral arteries as soon as the salinity was lowered. There was then a transient increase in anterolateral flow rates followed by a subsequent drop in flows between 8 and 9 h (Fig. 7E; ANOVA, F=4.06, P<0.01). Pre-treatment levels were regained after 2 h in 100%SW. Although there was an apparent decrease in flows through the hepatic arteries (Fig. 7F) in 65%SW with a recovery in 100%SW, no statistical significance could be demonstrated (ANOVA, F=1.71, P>0.05). Flows through the posterior aorta varied between 0.32 and 0.39 ml min^-1 in control conditions (Fig. 7G). There was no significant change during low salinity exposure. Upon return to 100%SW flow rates through the posterior aorta reached 0.59±0.06 ml min^-1; these were significantly elevated over pre-treatment and treatment conditions (ANOVA, F=4.7, P<0.001). The sternal artery received between 7.1 and 9.4 ml min^-1. Flow rates through this artery only decreased significantly after 4 h in low salinity(Fig. 7H; ANOVA, F=2.23, P<0.01). Pre-treatment levels were regained after 3 h exposure to 100%SW.

Ventilation rate was maintained between 76.4 and 78.8 beats min^-1 in control conditions. Apart from a slight increase 4 h after initiation of 65%SW (Fisher's LSD, P<0.01), there was no change in response to low salinity (Fig. 8A). After 3 h recovery in 100%SW ventilation rate had increased significantly reaching 88.8±3.6 beats min^-1 (ANOVA, F=4.29, P<0.001). These levels were maintained for the remainder of the recovery period. Oxygen uptake varied between 42.2 and 45.8 mg O₂ kg^-1 h^-1 in control conditions(Fig. 8B). There was a decrease in oxygen uptake when 65%SW was administered, falling to 19.75±4.4 ml min^-1 (ANOVA, F=8.82, P<0.001). However,within 1 h pre-treatment levels were regained. There was a further increase in oxygen uptake after 1 h in 100%SW (Fishers LSD, P<0.001) and it remained elevated for the duration of the recovery period.

The practice of starving animals prior to experimentation, to ensure they are in a similar metabolic state, may not be indicative of what happens in nature. Recent work has shown that feeding and subsequent assimilation of nutrients can alter the physiological responses of aquatic organisms to environmental perturbations (Legeay and Massabuau, 1999; Legeay and Massabuau, 2000; Whiteley et al., 2001; Robertson et al.,2002; Mente, 2003; Mente et al., 2003; McGaw, 2005; McGaw, 2006a). Animals may be able to prioritize or sum responses in order to balance the demands of physiological systems (Bennett and Hicks,2001; Hicks and Bennett,2004; McGaw,2005). However, the added cost of digestion in physiologically stressful environments can result in increased mortality for some crustaceans(Legeay and Massabuau, 2000; McGaw, 2006a).

Cancer gracilis exhibited the typical crustacean response to feeding; oxygen uptake increased immediately and remained elevated for over 48 h (Fig. 2B). The large initial increase was the result of increased activity during food handling and mechanical breakdown in the foregut [apparent specific dynamic action(Carefoot, 1990)]. The following longer term elevation would represent cellular protein synthesis(Houlihan et al., 1990; Mente, 2003). The cardiac and ventilatory responses associated with feeding and subsequent digestion in Cancer gracilis were similar to, but not as pronounced as those recorded for other species (McGaw and Reiber, 2000; McGaw,2005). There was, however, a large increase in sternal artery flow during feeding in Cancer gracilis(Fig. 1H). Although some of the apparent specific dynamic action associated with feeding activity was reduced by feeding the crabs via a catheter, they still moved the mouthparts,which are supplied via the sternal arterial system(McGaw and Reiber, 2002).

Essentially the effects of feeding opposed those observed during low salinity exposure. The decreases in cardiac function and haemolymph flow rates in unfed Cancer gracilis in low salinity (Figs 3, 4) were a result of behavioural adjustments (D. L. Curtis, E. K. Jensen and I. J. McGaw, manuscript submitted for publication). Cancer gracilis became quiescent, closing its mouthparts and retracting the antennae as soon as the salinity started to decrease. This closure response isolates the branchial chambers from the surrounding low salinity water (Sugarman et al., 1983). Because of the rapid isolation response as well as diffusive ion loss into a closed area, the water in the branchial chamber is held at a higher osmolality than the surrounding water (D. L. Curtis, E. K. Jensen and I. J. McGaw, manuscript submitted for publication). At the same time, the decreased cardiac output would result in a higher haemolymph residence time in the gills, which reduces the average exchange gradient for inward movement of water and diffusive ion loss(Cornell, 1973; Hume and Berlind, 1976). The sealed chamber and a decreased blood flow through the gills resulted in a reduction in oxygen uptake (Fig. 4B). These periods of `breath holding' can only be carried out for a short time before oxygen reserves are depleted, forcing opening and a subsequent increase in oxygen uptake (D. L. Curtis, E. K. Jensen and I. J. McGaw, manuscript submitted for publication). A decrease in cardiac output would also slow blood flow to the tissues. Low salinity is known to cause an increased haemolymph oxygen binding affinity in Carcinus maenas(Truchot, 1973), therefore a slowing of haemolymph flow at the tissue level could be beneficial in increasing oxygen extraction from the circulating haemolymph(Larimer, 1964).

It has been reported that organisms respond to a dilution of the medium by exhibiting an increase, a decrease or no change in respiration levels(Kinne, 1964). It has been suggested that euryhaline organisms show an increase and stenohaline organisms exhibit a decrease in respiratory and cardiac parameters, but in the past this had been difficult to substantiate(Wheatly, 1988). With the increase in respiratory and cardiovascular studies, a pattern is now emerging,to which many decapod crustaceans conform. The efficient osmoregulators remain active, increasing cardiac and respiratory parameters(Dehnel, 1960; King, 1965; Engel et al., 1975; Hume and Berlind, 1976; Cumberlidge and Uglow, 1977; Taylor, 1977; Spaargaren, 1982; Guerin and Stickle, 1992; McGaw and Reiber, 1998). Weaker regulators tend to become inactive(Sugarman et al., 1983; McGaw et al., 1999) and show no change in oxygen uptake (Brown and Terwilliger, 1999) and mixed cardiac responses(McGaw and McMahon, 1996; McGaw and McMahon, 2003; McGaw, 2006a; Dufort et al., 2001). Osmoconformers show a decrease in activity(McGaw et al., 1999) (D. L. Curtis, E. K. Jensen and I. J. McGaw, manuscript submitted for publication)and cardiac and respiratory parameters (Figs 3, 4)(Spaargaren, 1973; Cornell, 1973; Cornell, 1974; Cornell, 1979; Savant and Amte, 1995).

In the feeding-salinity experiments crabs were fed in 100%SW and then exposed to low salinity. This protocol would simulate crabs feeding over high tide (and highest salinity) followed by exposure to decreased salinity during low tide. The physiological responses of actual feeding during low salinity exposure were not investigated because Cancer gracilis did not feed in salinities below 80%SW (I.J.M., personal observation). When low salinity followed feeding, instead of a noticeable decrease, cardiac parameters and haemolymph flows remained stable for an extended period or even increased(Figs 5, 7). Differential changes in oxygen uptake were not as obvious; the crabs still exhibited the same isolation response. However, the drop in oxygen uptake was not as pronounced and because of the increased oxygen demand due to digestion, was of shorter duration compared with starved animals. This was followed by a rapid return to feeding levels. This pattern was similar 3 h after feeding when food has just started to enter the midgut and 24 h after feeding when the gut is cleared and protein synthesis is well underway (McGaw,2006b). Our recent work suggests that crabs may be able to suspend the specific dynamic action if they encounter hyposaline environments immediately after a meal (Curtis and McGaw, 2006). However, intracellular digestion can start within 2 h after ingestion of a meal (Houlihan et al., 1990; Mente,2003; Mente et al.,2003). Therefore, even at 3 h, intracellular digestion may have already started in Cancer gracilis and it would have to adjust its physiological responses accordingly in order to meet the greater metabolic demand.

Cancer gracilis can slow food processing in the gut and can even regurgitate food from the foregut in low salinity(McGaw, 2006b). This decrease in food processing is paralleled by a decrease haemolymph flows to the digestive gland via the hepatic arteries(Fig. 5F). These responses may spare energy for other systems. However, Cancer gracilis cannot halt digestive processes completely. Thus, the results from the present study suggest a prioritization of cardiac and respiratory responses associated with digestive events (McGaw and Reiber,2000; McGaw,2005). The prioritization of digestive events is in contrast to the weak regulator, Cancer magister, which tends to prioritize physiological responses to low salinity. However, postprandial Cancer magister exhibit a higher mortality rate in low salinity(McGaw, 2006a). This probably occurs because low salinity increases haemolymph oxygen binding affinity(Truchot, 1973), slowing the rate at which oxygen is offloaded to the tissues at a time when its use is enhanced by protein synthesis (Legeay and Massabuau, 2000). Because no differential mortality occurred for postprandial Cancer gracilis in low salinity (I.J.M., unpublished observation), the increases in cardiac and respiratory parameters observed in postprandial crabs in low salinity may optimize oxygen delivery to the tissues.

There was a pronounced overshoot in cardiac and ventilatory parameters when crabs were returned to 100%SW, following low salinity exposure. It is interesting to note that increases in cardiac function and haemolymph flow rates were greater in postprandial crabs when returned to 100%SW. This rapid increase in cardiac function and haemolymph flows was not due to an increase in mechanical digestion, because it takes several hours for gastric processing to resume once 100%SW is restored (McGaw,2006b). More likely this increase would help repay an oxygen debt(Herried, 1980) caused by the extra demand of intracellular digestion and subsequent protein synthesis(Mente, 2003).

In the natural environment of Cancer gracilis, low salinity episodes are usually associated with tidal changes and only last a few hours(D. L. Curtis, E. K. Jensen and I. J. McGaw, manuscript submitted for publication). Therefore, the observed physiological responses would be adequate to allow postprandial Cancer gracilis to cope with acute hyposaline exposure. Nevertheless, the present study has shown the nutritional status of an animal can alter physiological responses. Consequently, previous`controlled' laboratory experiments, where animals were starved prior to experimentation, may not be wholly representative of physiological processes occurring in nature. This underscores the importance of studying physiological responses at the whole organism level and across the range of environmental conditions under which they operate.

This work was carried out during a UNLV research sabbatical. I would like to thank the Director and staff of the Bamfield Marine Sciences Centre for use of facilities. Supported by a grant from the National Science Foundation IBN#0313765.