Oxygen consumption remains stable while ammonia excretion is reduced upon short time exposure to high salinity in Macrobrachium acanthurus (Caridae: Palaemonidae), a recent freshwater colonizer

Palaemonid shrimps occur in the tropical and temperate regions of South America and the Indo-Pacific, in brackish/freshwater habitats, and marine coastal areas. They form a clade that recently (i.e., ~30 mya) invaded freshwater, and one included genus, Macrobrachium Bate, 1868, is especially successful in limnic habitats. Adult Macrobrachium acanthurus (Wiegmann, 1836) dwell in coastal freshwaters, have diadromous habit, and need brackish water to develop. Thus, they are widely recognized as euryhaline. Here we test how this species responds to a short-term exposure to increased salinity. We hypothesized that abrupt exposure to high salinity would result in reduced gill ventilation/perfusion and decreased oxygen consumption. Shrimps were subjected to control (0 psu) and experimental salinities (10, 20, 30 psu), for four and eight hours (n = 8 in each group). The water in the experimental containers was saturated with oxygen before the beginning of the experiment; aeration was interrupted before placing the shrimp in the experimental container. Dissolved oxygen (DO), ammonia concentration, and pH were measured from the aquaria water, at the start and end of each experiment. After exposure, the shrimp’s hemolymph was sampled for lactate and osmolality assays. Muscle tissue was sampled for hydration content (Muscle Water Content, MWC). Oxygen consumption was not reduced and hemolymph lactate did not increase with increased salinity. The pH of the water decreased with time, under all conditions. Ammonia excretion decreased with increased salinity. Hemolymph osmolality and MWC remained stable at 10 and 20 psu, but osmolality increased (~50%) and MWC decreased (~4%) at 30 psu. The expected reduction in oxygen consumption was not observed. This shrimp is able to tolerate significant changes in water salt concentrations for a few hours by keeping its metabolism in aerobic mode, and putatively shutting down branchial salt uptake to avoid massive salt load, thus remaining strongly hyposmotic. Aerobic metabolism may be involved in the maintainance of cell volume, concomitant with reduced protein/aminoacid catabolism upon increase in salinity. More studies should be conducted to broaden our knowledge on palaemonid hyporegulation.


INTRODUCTION
Palaemonid shrimps have a wide global distribution.They occur in a great variety of aquatic environments, from seawater up to full freshwater (Augusto et al. 2009, Anger 2013).The family Palaemonidae has a marine origin and passed through various independent events of freshwater invasion (Murphy and Austin 2005, Ashelby et al. 2012, McNamara and Faria 2012, Anger 2013).These events have occurred quite recently, ~30 mya (Tertiary), and are still happening (Murphy and Austin 2005, Augusto et al. 2007a, b, 2009, Pileggi and Mantelatto 2010, Collins et al. 2011, McNamara and Faria 2012).The family comprises about 116 genera.Among these, Macrobrachium Bate, 1868 was the most successful colonizer of freshwater and estuarine waters, showing a wide geographical distribution (Anger 2013).One species in particular, commonly referred to as "giant freshwater prawn," Macrobrachium rosenbergii (De Man, 1879) has large economic importance globally (Bond-Buckup andBuckup 1989, Valenti et al. 1989).
The fact that freshwater Palaemonid shrimps have recently transitioned from saline waters to more dilute waters renders them quite tolerant to increased salinity (Freire et al. 2008a(Freire et al. , 2013)).Freshwater Palaemonid shrimp adults -especially the more coastal and diadromous species -those whose larvae depend on brackish waters for their proper development (Charmantier 1998, Anger 2003, McNamara and Faria 2012), are particularly euryhaline (McNamara 1987).
Estuarine, but especially freshwater crustaceans, are good hyper-osmoregulators, that is, they keep steep osmotic and ionic gradients with respect to the surrounding water, aided by the low permeability of their cuticle (Péqueux 1995, Freire et al. 2003, 2008a, b).Palaemonid shrimps also follow this pattern (Freire et al. 2003, Murphy and Austin 2005, Augusto et al. 2009, Boudour-Boucheker et al. 2013).It is during ecdysis that cell volume may be challenged in these shrimps, as their internal medium fluctuates (compared to the intermoult period), diluting strongly in fresh waters.However, crustaceans in general, palaemonid shrimps in particular, can face this challenge by regulating cell/tissue hydration quite efficiently (see Freire et al. 2013).These hyper-regulation mechanisms, one of the most prominent features of freshwater crustaceans (second to the cuticle) have been frequently studied in Palaemonidae (e.g., Freire et al. 2008b, McNamara andFaria 2012).Hyporegulation mechanisms, by contrast, have been much less investigated and remain elusive (Freire et al. 2008b, McNamara andFaria 2012).
Salinity challenges can also result in changes in metabolic responses.For instance, a decrease in the respiratory rate of the marine shrimps Marsupenaeus (Penaeus) japonicus (Bate, 1888) (Setiarto et al. 2004), and Litopenaeus (Penaeus) setiferus (Linnaeus, 1767) (Rosas et al. 1999), when exposed to a decrease in salinity, was observed.Some palaemonid shrimps exposed to salinity increases, for instance Macrobrachium heterochirus (Wiegmann, 1836) and Macrobrachium potiuna (Müller, 1880), experience a decrease in metabolic rates, while the diadromous Macrobrachium acanthurus (Wiegmann, 1836) and Macrobrachium olfersii (Wiegmann, 1836) experience a peak in their metabolism-salinity curves ("dome-shaped curve") close to their isosmotic point, ~21 psu (Moreira et al. 1983).In another diadromous shrimp, Macrobrachium amazonicum (Heller, 1862), oxygen consumption was lower in freshwater than in 18 psu in zoea II, and was higher in freshwater than in 12 and 18 psu in zoea V (Mazzarelli et al. 2015).The freshwater shrimp Macrobrachium tuxtlaense Villalobos and Alvarez, 1999, when exposed to increased salinities up to 30 psu, has shown an increase in oxygen consumption in 5 and 10 psu, and a decrease in this parameter in the other salinities, with respect to the control salinity, fresh water (Ordiano et al. 2005).With this variability in the metabolic response of shrimps in the background, the aim of this study was to test whether a short term exposure to increased salinity would result in a "shut down"of oxygen uptake, a putative "escape response", potentially activating anaerobic metabolism and lactate production in this diadromous palaemonid.

MATERIAL AND METHODS
Specimens of M. acanthurus were bought from local fishermen from Rio dos Barrancos (25°36'32.0"S,48°24'02.5"W),municipality of Pontal do Paraná, Paraná, Brazil, who sell them as live bait.Shrimps were transported to the laboratory for approximately two hours, in plastic gallons with constant aeration.The animals were acclimated for about five days in 35 liters aquaria with fresh water (double filtered tap water, charcoal and cellulose filters), in temperature of 20±1°C, constant aeration, and natural photoperiod (~12 h light: 12 h dark).Some ions were assayed in our tap water (mean±standard deviation, in mM, n = 6 for all): chloride 0.23 ± 0.29; magnesium: 0.16 ± 0.06; sodium 4.67 ± 1.94; potassium 0.57 ± 0.20, and osmolality of 26.2 ± 4.3 mOsm/kg H 2 O. Shrimps were fed fragments of fish fillet on alternate days.Shrimps (5.2 ± 0.7 cm, n = 64) were individually subjected to salinities 0 (control), 10, 20, or 30 psu, for 4 or 8 hours (n = 8 for each coupled condition of salinity x time), in 250 ml containers, water temperature of 21.1 ± 0.05 °C.Saline waters were obtained through proportional mixture of filtered tap water with natural sea water.The experiments were conducted without aeration, in order to allow the determination of oxygen consumption by the shrimp, but the initial water was saturated with oxygen, through overnight aeration, before the start of the experiments (initial oxygen concentration of 7.49 ± 1.19 mg/l for 0 psu, 7.36 ± 0.60 mg/l for 10 psu, 6.73 ± 0.15 mg/l for 20 psu, and 6.55 ± 0.15 mg/l for 30 psu, n = 16 for each salinity).The following water parameters were analyzed at the initial (before placing the shrimp in the container) and final (after removing the shrimp from the container) times of exposure: dissolved oxygen, pH, and ammonia.Differences between the initial and final concentrations of dissolved oxygen and ammonia represented, respectively, oxygen consumption and ammonia excretion by the shrimp (N-NH 3 ).There was essentially no ammonia in the water at the beginning of the experiments (0.008 ± 0.003 mg/l of N-NH 3 , n = 70 samples).Experiments were also conducted in containers with water but without animals, as blanks for water parameters (n = 6 for each experimental condition, yielding a total of 72 containers).
After the stipulated times of exposure, the animals were cryoanesthetized (covered with ground ice) for about 1 minute, until fully immobile.Then, hemolymph samples were collected through cardiac puncture, with a micropipette inserted under the exosqueleton, for determinations of lactate and osmolality.Finally, the exosqueleton was removed, and a fragment of abdominal muscle was collected for determination of water content.
Other individuals were subjected to the same protocol of salinity increase (n = 3 for each condition of salinity x time), to evaluate whether shrimps were ventilating their gills equally in high salinity media, as they do in their habitat, fresh water.The hypothesis was that their gills would get stained from the dye added to the water, after some minutes of exposure, from gill ventilation.At the end of the experimental exposures to high salinity, five drops of 1% methylene blue were added to each of the 250 ml containers.Shrimps were maintained in these conditions for 5 minutes, after which they were removed from the containers and had both sides of their cephalotorax photographed, with focus on their gills.The intensity of the blue staining of their gills was qualitatively evaluated.The same procedure with the dye was conducted for control shrimps in fresh water (see Suppl.material 1).
The levels of dissolved oxygen were detected in the water through an oxymeter (YSI model 55, USA).Water pH was determined using a bench pHmeter (inoLAB pH Level 1WTW, Germany).The concentration of ammonia was assayed through colorimetric commercial kits (Alfakit, Brazil), and absorbance was read at 630 nm (Spectrophotometer Ultrospec 2100 PRO Amersham Pharmacia biotech, Sweden).
Hemolymph lactate was assayed through colorimetric commercial kits (Labtest, Brazil), with absorbance read at 550 nm.Hemolymph osmolality was determined using a vapour pressure osmometer in undiluted samples (Vapro 5520, Wescor, USA).For the determination of muscle water content, tissue fragments were weighed (wet weight, analytical balance Bioprecisa FA2104N, Brazil, precision of 0.1 mg), dried in an oven at 60 °C for 24 hours, then weighed again (dry weight).The difference between wet and dry weights, as a percentage, represents the muscle water content, or its hydration.
Two-way ANOVAs (factors were salinity and time) with post hoc tests of Holm-Sidak were conducted for each of the following parameters: oxygen consumption, lactate, osmolality, and muscle water content.Initial and final pH values did not pass the normality and equal variance tests.These data were transformed to meet the requirements of the parametric two way ANOVA.Ammonia values could not be normalized, and for this reason they were treated differently.Two non-parametric (Kruskal-Wallis) "one-way-ANOVAs" were conducted, one for 4 hours, one for 8 hours.The respective values of each salinity, 4 vs 8 hours, were compared using t-tests.The initial versus final values of dissolved O 2 , and pH in the water of containers were compared through paired t-tests for each experimental condition.Pearson correlations were performed for factors salinity, oxygen consumption, lactate, osmolality, muscle water content, excreted ammonia, final pH, and total length.The adopted significance level was 0.05.

Experimental blanks for water parameters
The initial and final water parameters (O 2 , pH, and NH 3 ) in the blanks, vials without any shrimp -for all salinities and times of exposure -are shown in Table 1.

Water oxygen consumption and lactate concentration in the hemolymph
The two-way ANOVA revealed that time (F = 14.1, p < 0.001) and salinity (F = 2.9, p = 0.043), but not their interaction (F = 0.66, p = 0.58) affected water oxygen consumption in M. acanthurus.The initial oxygen concentration in the water was always higher than the final concentration, for all salinities and times of exposure, indicating oxygen consumption by the shrimp (Fig. 1).However, consumption, as quantified by the difference between initial and final oxygen levels in the water of the container, was stable and did not vary among the experimental treatments and controls.The two-way ANOVA revealed that time (F = 4.3, p = 0.042), but not salinity (F = 0.068, p = 0.977), or their interaction (F = 0.15, p = 0.93) affected hemolymph lactate in M. acanthurus.Production of lactate by the shrimp, measured from its hemolymph, was essentially invariable in the several experimental conditions (Fig. 2).

pH and excreted ammonia
The two-way ANOVA on initial pH values revealed no effect of time (F = 0.79, p = 0.38), but an effect of salinity (F = 132, p < 0.001), and no interaction between time and salinity (F = 0.82, p = 0.49).The final pH of the water , according to the two-way ANOVA, was affected by time (F = 19.5, p < 0.001), and salinity (F = 56.3,p < 0.001), but not by their interaction (F = 1.100, p = 0.36).The pH of the water was always higher at the start (initial) of the experiment than after 4 and 8 hours (final, Fig. 3).The Kruskal-Wallis tests revealed that salinity had an effect on water ammonia concentrations after 4 hours (p < 0.001), and after 8 hours (p < 0.001).The amount of excreted Table 1.Initial and final concentrations of oxygen and ammonia, and values of pH in the water of "blank" containers, without any shrimp, in salinities 0 (control), 10, 20, and 30 psu, for 4 and 8 hours of exposure (n = 6 for each group).(2) of M. acanthurus exposed to salinities 0 (control), 10, 20, and 30 psu for 4 (white symbols) and 8 hours (black symbols).Values near to the lines represent oxygen consumption (mean ± std dev) as a percentage of the initial oxygen concentration (considered 100%, as a reference value).(#) Initial and final levels of dissolved oxygen are different.There were no significant differences between salinities or times (4 and 8 hours) for oxygen consumption and lactate concentration.ammonia (N-NH 3 ) decreased with increasing salinity, with the highest value in 0 and 10 psu, and the lowest values in 20 and 30 psu.The effect of time on the amount of excreted ammonia was noted in the controls in 0 only, with higher ammonia levels measured in the water after 8 hours than after 4 hours (Fig. 3).

Osmolality of the hemolymph and muscle water content
The two-way ANOVA revealed that time (F = 12.9, p < 0.001) and salinity (F = 79.8,p < 0.001), and their interaction (F = 7.3, p < 0.001) had an effect on the osmolality of the hemo-lymph of M. acanthurus.It increased by 20 psu, and further by 30 psu with respect to the control (0 psu), after 4 and 8 hours.An effect of time was noted at 30 psu: osmolality was higher after 8 hours than after 4 hours (Fig. 4).Coherently, muscle water content showed the opposite trend, that is, it decreased as salinity increased.The two way ANOVA revealed that salinity (F = 25.3, p < 0.001) and the interaction between salinity and time (F = 2.95, p = 0.041) affected the hydration of the muscle, but time alone did not (F = 2.3, p = 0.14).In the highest salinity, 30 psu, muscle water content after 8 hours was lower than after 4 hours, matching the raised osmolality of the hemolymph (Fig. 5).

DISCUSSION
Water dissolved oxygen (DO) was consumed by M. acanthurus during the experiments, and was consistently detected by our assay method, which is evidenced by the reduction in water DO.No decrease in water DO was detected under the same experimental conditions (volume, temperature, previous Figures 4-5.Hemolymph osmolality (4) and muscle water content (5) of M. acanthurus exposed to salinities 0 (control), 10, 20, and 30 psu for 4 (white circles) and 8 hours (black circles).Dashed line represents water expected values, from the relationship 1 psu = 30 mOsm/kg H 2 O, short horizontal lines indicate value of calculated water osmolality for the tested salinities.Different letters mean differences between salinities within each time of exposure.* = value for 4 hours is different from value for 8 hours within a same salinity.
DO saturation protocol, DO electrode), but without a shrimp in the vial ("blanks"), supporting the conclusion that oxygen was indeed consumed by the shrimp.Unexpectedly, salinity did not affect oxygen consumption by M. acanthurus after 4 or 8 hours of exposure; the correlation between these two variables was significant, but weak.
The relationship between salinity and oxygen consumption in shrimps is rather complex and variable.In the marine palaemonid Palaemon serratus (Pennant, 1777) no change in oxygen consumption between salinities 34 and 15 psu (Salvato et al. 2001) was observed.In contrast, in the marine shrimp L. (P.) setiferus the effect of salinity on the rate of oxygen consumption was influenced by the developmental stage of the individual: post-larvae PL10-PL15 showed the highest oxygen consumption at 10 psu, while post-larvae PL15-PL21 showed the highest oxygen consumption at 40 psu (salinity ranged from 5 to 40 psu - Rosas et al. 1999).
Importantly, in the study cited immediately above (Rosas et al. 1999), oxygen consumption in penaeid shrimps was measured in these cited salinity levels after 120-264 hours of exposure them.In addition, penaeid marine shrimps are hyper-hypo-regulators (Péqueux 1995, Freire et al. 2008b), as opposed to palaemonid shrimps, which are essentially hyper-regulators, as already mentioned.In the results of an experiment with M. amazonicum at different ontogenetic stages (zoea I, II, V, and IX) exposed to different salinities (0.5, 6, 12 or 18 psu), there was a greater consumption of oxygen at 0.5 psu in the zoea stage V, probably due to the great amounts of energy required for the active transport of salts through the epithelia (Mazzarelli et al. 2015).A similar pattern was observed in M. tuxtlaense, an strictly freshwater prawn exposed to a salinity gradient (0, 5, 10, 15, 20, 25, and 30 psu): higher rates of oxygen consumption in the shrimp M. tuxtlaense were observed at low salinities (0, 5 and 10 psu), to account for hyper-regulation of the osmolality of the hemolymph (Ordiano 2005).In the results of Moreira et al. (1983) working with M. acanthurus, the oxygen consumption curve after 24 hours of exposure was dome-shaped, peaking at 21 psu.These results show that the metabolic response to variations in salinity is indeed variable in these decapod crustaceans, is dependent on the osmoregulatory history and strategy of the species, and is also time-dependent.
Consistent with the results on oxygen consumption rates, the concentration of hemolymph lactate in M. acanthurus remained constant, they did not change when salinity increased.The hypothesis here was that increased salinity would lead to a reduction in gill perfusion and, consequently, reduced oxygen consumption.Reduced oxygen consumption would result in anaerobic metabolism and lead to increased levels of hemolymph lactate (Booth et al. 1982).In many decapod crustaceans, lactate is the main product of metabolism when there is a hypoxic condition (Bridges and Brand 1980, Taylor and Spicer 1989, Maciel et al. 2008).These shrimps continued to take up oxygen from the water, and did not enter into functional hypoxia.Had they entered into intense anaerobic metabolism, increased hemolymph lactate would have been detected.However, metabolic carbon dioxide was produced and released through the gills, causing water acidification (Henry and Wheatly 1992), which was detected in our results.
Ammonia release decreased with increased salinity (strong and significant negative Pearson correlation).One possible factor that could at least partially account for this inverse relationship is the fact that NH 3 can be excreted as NH 4 + , especially in acidic water, and in animals with acidosis, replacing K + in the Na + / K + -ATPase (e.g., Claiborne et al. 1982, Wall 1995, Furriel et al. 2004).If Na + /K + -ATPase activity is reduced with increased salinity in this shrimp (see Maraschi et al., 2015), then NH 4 + transport from the hemolymph also putatively decreases, leading to less ammonia in the water, as observed here.Alternatively, ammonia may also be transported and eliminated through Rh-proteins (see Weirauch et al. 2009).
This freshwater shrimp strongly hyper-regulates in freshwater (gradient of +400 mOsm/kg H 2 O), its natural habitat in the adult phase, and continues to show hyper-osmotic hemolymph after 4-8 hours in 10 psu (+100 mOsm/kg H 2 O).However, after 4-8 hours in 20 or 30 psu, although there is some increase in hemolymph osmolality (strong positive correlation between salinity and hemolymph osmolality), it becomes hyposmotic to the water at -150 and -300 mOsm/kg H 2 O, respectively.
When there is a significant salt load, for instance 10, 20, 30 psu, what happens to the salt uptake system of the gills of freshwater shrimps, which normally steeply absorb salt from freshwater?The first hypothesis that can explain the relative osmotic stability of the hemolymph is that gill ventilation/ perfusion would drastically decrease, especially when the exposure is short (up to a few hours).When this happens, consumption of oxygen from the water also decreases.Such decrease in oxygen consumption, however, was not observed in our data.In fact, under all experimental conditions tested here, when a vital dye (methylene blue) was pipetted next to the shrimp, and the branchial chamber and gills were observed under a stereomicroscope, the gills were stained blue (data not shown, see Suppl.material 1).There was great variability in the resulting blue color of the gills of the shrimps.The idea behind this experiment using the blue stain was that, if the shrimp reduced its gill ventilation upon increased water salinity, oxygen consumption from the water would be reduced and hence its gills would remain clear, whitish, when tested.Apparently, however, compatible with the oxygen consumption data, M. acanthurus apparently perfuses its gills even under severe salinity increase/stress.
When salinity rises beyond the organisms' homeostatic range of osmoregulation, the osmolality of the hemolymph increases with respect to values in lower salinities but still remains below the osmolality of the water.And this, in turn, beyond a certain limit, leads to an inability to control tissue hydration and volume.The water content in the muscle of M. acanthurus was inversely proportional to salinity (strong negative Pearson correlation), and was maintained within a narrow range of variation, with a decrease of ~4-5% in 30 psu with respect to the control (0 psu), after 4-8 hours of exposure.Conversely, at 30 psu, hemolymph osmolality increased by ~50%.This means that, even when the hemolymph experienced a great increase in osmotic concentration, the hydration of the muscle varied very little, indicating that this tissue has high capacity to regulate water concentrations.A similar result was observed in M. acanthurus at 30 psu (Foster et al. 2010).The great ability of tissue to maitain its water concentration was also documented for M. acanthurus and the other palaemonids M. potiuna and P. pandaliformis, through an "in vitro" experiment in which tissues were exposed to hypo-and hyperosmotic saline solutions that corresponded to a 50% change with respect to the isosmotic control.In this study, the hydration of the shrimp tissues varied in only about 10% (Freire et al. 2013).This high capacity to maintain tissue water levels is in part responsibe for the euryhalinity of M. acanthurus, which, throughout its life cycle, switches between freshwater and brackish water (Freire et al. 2008a(Freire et al. , 2013) ) The maintenance of tissue hydration happens through the regulation of the flux of inorganic ions and concentration of aminoacids or other nitrogenous compounds in the tissues or body fluids (e.g., Pierce 1982, Gilles 1987).Under hyperosmotic challenges, osmolyte concentrations increase in tissues, as already shown for the crustaceans M. amazonicum, M. olfersii, Dilocarcinus pagei (Augusto et al. 2007a, b).As a consequence, aminoacid catabolism should decrease, which is compatible with our results.Thus, the decrease in ammonia excretion observed in our study under hypersaline challenges probably means that there is a mechanism to retain aminoacid or nitrogenous compounds, which allows the maintenance of tissue hydration in high salinities.Compatible with this idea, excretion of ammonia in our results was inversely proportional to salinity (as pointed by the Pearson correlation).The role of aminoacids in the maintenance of tissue hydration has been documented in recent freshwater invaders (Augusto et al. 2007a,b).In summary, when there is an increase in salinity, metabolic energy is routed to controlling extracellular and intracellular homeostasis, by shutting down branchial salt uptake and reducing protein/ aminoacid catabolism.The palaemonid shrimp studied here, a recent freshwater invader, does not display "avoidance" or "escape" response when faced with severe salt challenges, even considering its strong and effective apparatus to perform salt uptake from freshwater.An avoidance behaviour would result in reduced oxygen consumption from the water.Rather, although certainly reducing salt uptake, this shrimp maintains its hemolymph hyposmotic with respect to the ambient water at 20-30 psu for 4-8 hours, during which it uses more oxygen to control water tissue levels .Additional studies are needed to elucidate the limits, degrees, and mechanisms of hypo-regulation or hypo-conformation in palaemonid shrimps.

Supplementary material 1
Figure S1.Cephalothorax of Macrobrachium acanthurus exposed to salinities 0 (control), 10, 20, and 30 psu for 4 or 8 hours.Authors: Carolina A. Freire, Leonardo de P. Rios, Eloísa P. Giareta, Giovanna C. Castellano Data type: JPEG image file Explanation note: The aquarium water was dyed with methylene blue, coloring the gills and gills chamber according to their ventilatory perfusion.After 5 minutes, shrimps were removed and photographed.An "uncolored" shrimp was included for reference, not exposed to the dye.Despite the variability in the response, it is possible to detect the stain in shrimps exposed to all experimental conditions.

Figure 3 .
Figure3.Initial (circles) and final (squares) pH and excreted ammonia (N-NH 3 ) in water of M. acanthurus exposed to salinities 0 (control), 10, 20, and 30 psu for 4 (white symbols) and 8 hours (black symbols).Different letters mean differences between salinities within each time of exposure.* = value for 4 hours is different from value for 8 hours within a same salinity, # = initial and final values of water pH are different.