Does physiological tolerance to acute hypoxia and salinity change explain ecological niche in two intertidal crab species?

Lay summary In response to acute salinity change or hypoxia, Hemigrapsus crenulatus exhibited better maintenance of osmoregulatory and cardiovascular function than Hemigrapsus sexdentatus. The greater physiological resilience of H. crenulatus is consistent with its habitation of an ecological niche that promotes greater exposure to environmental stressors than that of H. sexdentatus.


Introduction
Intertidal environments are in constant flux. Over the course of a tidal cycle, an aquatic organism living in an intertidal setting may experience changes in water availability, variable water salinity, temperature oscillations and fluctuations in dissolved oxygen (DO). For example, tidal pool DO has been shown to range from 10 to 955 μmol l −1 (∼0.8-87 kPa) over the course of 6 h (Legrand et al., 2018), whilst salinity in estuaries can range from near zero to hypersaline depending on the magnitude of freshwater influx (Montagna et al., 2018). This instability of environmental physicochemical factors challenges organism homeostasis. Animals that live in these settings must employ physiological strategies that allow them to maintain function despite rapid environmental fluctuation and/or enact behaviours that minimize their exposure to these factors.
Crabs are common inhabitants of intertidal zones and employ a number of physiological mechanisms that enable them to withstand fluctuations in environmental variables such as salinity and DO. Osmoregulating crab species maintain haemolymph osmolality and ion concentrations through the co-ordinated actions of epithelial transporters (Péqueux, 1995). However, epithelial transport may incur a metabolic cost that can be manifested as an increase in oxygen consumption (Shinji et al., 2009). Altered energy demands may also be reflected in changes in cardiovascular parameters. For example, tachycardia is commonly reported in crabs exposed to dilute salinities (Hume and Berlind, 1976;Taylor, 1977;McGaw and McMahon, 1996). This increase in heart rate may facilitate oxygen loading at the gills and thereby fuel the increased metabolic demands of osmoregulation (Taylor, 1977). In conditions of declining DO, bradycardia is frequently reported in crabs (Airriess and McMahon, 1994;McMahon, 2001;McGaw and McMahon, 2003), a response that slows haemolymph flow through the gills, facilitating oxygen transport under conditions where the oxygen uptake gradients are compromised relative to normoxic waters (McMahon, 2001). However, at a certain critical oxygen partial pressure (PO 2 ), known as the P crit , regulation fails and the animal transitions from oxyregulation to oxyconformation (Yeager and Ultsch, 1989). This is associated with a greater reliance on anaerobic metabolism. Critically, because of the metabolic costs associated with osmoregulation and the reduced energy provided by anaerobic metabolism, a decrease in DO might ultimately compromise the ability of a crab to maintain extracellular fluid osmolality and ion concentrations (Lucu and Ziegler, 2017).
On New Zealand coastlines, two species of crabs belonging to the genus Hemigrapsus are regularly encountered. Hemigrapsus sexdentatus (Hilgendorf) (formerly H. edwardsi; McLay and Schubart, 2004), the purple rock crab, has a relatively wide ecological niche. It is mainly distributed on stony beaches, occupying the high to mid-tide zone (McLay, 1988). This species does not burrow, but instead may be found sheltering under stones. Notably, Hemigrapsus sexdentatus is often associated with freshwater inputs (Williams, 1969) and as a semi-terrestrial species, largely avoids fluctuations in seawater (SW) salinity associated with tidal rhythms. Conversely, Hemigrapsus crenulatus (Milne-Edwards), known as the hairy-handed crab, mostly occupies estuaries and lower intertidal zones. It is commonly associated with mud-flats and in these and other soft substrates, it will burrow (McLay, 1988). Sheltering in the burrow during a tidal cycle will expose the crab to a declining DO, whilst riverine inputs at low tide reduce ambient salinity. Both of these crabs are tolerant to variable salinities (Hicks, 1973) and are known to be good osmoregulators, maintaining haemolymph osmolality above SW osmolality in dilute waters (Taylor and Seneviratna, 2005;Lee et al., 2010;Urzúa and Urbina, 2017). To date, however, little is known regarding the tolerance of either of these species to low DO. Given the propensity of H. crenulatus to burrow into potentially anoxic, organic-rich, muddy sediments (McLay, 1988), this species would be predicted to be more hypoxia tolerant than H. sexdentatus.
The aim of the current study was to investigate the responses of H. crenulatus and H. sexdentatus to acute changes in DO and salinity, characteristic of those occurring during a tidal cycle. Our hypothesis was that differences in physiological capacity for homeostasis in response to environmental stressors such as low DO and reduced salinity, may contribute towards the distinct ecological niches of these two Hemigrapsus species on New Zealand coastlines and may ultimately determine their capacity to withstand anthropogenic pressures on their habitats.

Animal collection and maintenance
Male H. crenulatus {mean [±standard error of mean (SEM)] mass = 11 (±2) g} were collected from the Avon-Heathcote Estuary/Ihutai (43 • 33 S, 172 • 43 E), whilst male H. sexdentatus (33 ± 3 g) were collected from Waipara Beach (43 • 09 S, 172 • 48 E), both locations in the Canterbury province of New Zealand. Crabs were immediately transported to the University of Canterbury, where they were held in recirculating natural SW (salinity ∼ 35) maintained at 15 • C and subjected to a 12-h light:12-h dark photoperiod, for at least a week before experimentation. During acclimation to holding conditions, crabs were fed every other day on fresh mussels, although feeding was withheld 24-h prior to experimentation. All animal procedures were approved by the University of Canterbury Animal Ethics Committee.

Oxygen consumption and determination of critical PO 2
Oxygen consumption was measured using closed-boxed respirometry, via methods described previously for freshwater crayfish (Broughton et al., 2017) respirometer was placed in a 15 • C water bath and the respirometer filled with water at one of five salinities [0.7 (2% SW), 9 (25% SW), 18 (50% SW), 35 (100% SW) or 53 (150% SW)]. Low salinities were made by appropriate dilutions of natural SW collected from Lyttleton Harbour (43 • 36 S, 172 • 42 E) with distilled water, whilst the hypersaline water was made by adding artificial sea salt (Instant Ocean) to natural SW. Salinities were confirmed by vapour pressure osmometry (Wescor Vapro). After 10 min of vigorous aeration and circulation of the bathing medium through the respirometer, an individual crab (n = 6), was added into a chamber, which was then sealed and measurement of oxygen consumption commenced immediately thereafter (see Discussion). An indwelling oxygen electrode (Strathkelvin) facilitated continuous monitoring of respirometer PO 2 until it reached 1.1 kPa [6% of the initial, fully-oxygenated, water (18.7 kPa); ∼12 h]. These data were binned into 5 min intervals and the first 5-min period of each hour was used to plot the decline in PO 2 with time for each individual crab. Oxygen consumption rate (MO 2 ; μmol g −1 h −1 ) was calculated as follows: where a is the oxygen capacitance of water at 15 • C (in μmol l −1 kPa −1 ), PO 2 is the change in PO 2 in kPa, V is the volume of the respirometer (corrected for the presence of the crab), w is the mass of the crab in g and t is the time in h. Control respirometers (with no animal) were used to confirm that observed changes in PO 2 were not the result of factors such as the biological oxygen demand of the exposure water or respirometer failure. The relationship between MO 2 and PO 2 was graphed and a piecewise regression was performed using R statistical software to identify the transition point at which crabs change from oxyregulation to oxyconformation (i.e. the critical PO 2 or P crit ).

Acute salinity and hypoxia exposures
Individual crabs (n = 6 per species per treatment) were added directly to 450-ml Perspex chambers held within a recirculating water bath at 15 • C, containing waters that differed either in salinity (0.7, 9, 18, 35 or 53)  . Two sets of six crabs were used, one set for determination of heart rate and the other set for haemolymph sampling, which was thereafter analysed for osmolality and ions (sodium, potassium and chloride).
Heart rate was measured continuously using an infrared sensor attached non-invasively to the carapace, as described in Broughton et al. (2017). The attachment of the sensor housing was performed 2-3 days prior to experimentation. The mean heart rate over a 5 min period at the end of each hour of the exposure was recorded, excluding any intervals where the heart was not beating.
At the conclusion of the 6-h exposure, haemolymph was extracted from the infrabranchial sinus at the base of the walking legs. Haemolymph osmolality was determined using a vapour pressure osmometer (Wescor), sodium and potassium were assessed via flame photometry (Sherwood), whilst chloride was measured using a digital chloridometer (Labconco).

Statistical analyses
The two crab species differed significantly in mass (H. crenulatus = 11 g; H. sexdentatus = 33 g). Consequently, to avoid complications associated with mass differences, the responses of each crab species to salinity and DO were assessed independently (i.e. via one-way ANOVA) to examine the effects of experimental variables (i.e. DO, salinity) within a species. An initial two-way ANOVA showed that the heart rate of both species was not dependent on time (i.e. did not vary over the 6 h of exposure). Subsequently, one-way ANOVAs were conducted using the mean heart rate across the 6-h exposure for each individual.
For all analyses, tests of normality (Kolmogorov-Smirnov) and equality of variance (Levene's) were first conducted. When data were confirmed as parametric a one-way ANOVA was performed, followed by a post hoc Tukey's test. Nonparametric data were either log-transformed (effect of DO on H. sexdentatus heart rate) and then subjected to parametric ANOVA or interrogated via a non-parametric Kruskal-Wallis ANOVA, followed by a Dunn's post hoc test (effect of DO on H. crenulatus haemolymph potassium and chloride). For all analyses, an alpha value of 0.05 was considered significant. Throughout the manuscript, values are reported as means ± SEM.

Results
The normoxic MO 2 of H. crenulatus varied significantly as a function of exposure salinity (one-way ANOVA, P = 0.004; Fig. 1A). A fall in exposure salinity from full-strength SW (35) to 50% SW (18) resulted in a significant 1.5-fold increase in MO 2 . At lower salinities and in the hypersaline test condition, the MO 2 of H. crenulatus was unchanged relative to the control. This pattern was distinct from that observed for H. sexdentatus (Fig. 1B). In this species, there were no significant effects of salinity on MO 2 , although a trend towards increasing MO 2 with decreasing salinity could be observed (one-way ANOVA, P = 0.132).  Closed-box respirometry resulted in a decline in PO 2 as the crab consumed the oxygen in the respirometer (Fig. 2). Plotted values in this figure represent datapoints for all individual crabs within a given stressor level. In both crab species, in all water salinities, MO 2 remained relatively constant until a water PO 2 of ∼5.5-7.5 kPa. Thereafter, the capacity of the crab to regulate oxygen consumption failed and a rapid decline in MO 2 with falling water PO 2 was observed. Using values calculated for each individual and then averaged, the P crit was determined to be 5.9 ± 0.9 kPa for H. crenulatus in full-strength SW (i.e. salinity = 35). Under the same exposure conditions, the mean calculated P crit for H. sexdentatus was 7.6 ± 0.2 kPa. Values of P crit did not differ as a function of exposure water salinity for either species (one way ANOVAs; P = 0.37 for H. crenulatus, P = 0.08 for H. sexdentatus; data not shown).
During a 6-h exposure to a range of water salinities, the heart rate of H. crenulatus remained unchanged (oneway ANOVA, P = 0.92; Fig. 3A), ranging from 106 ± 11 to 121 ± 14 beats min −1 for salinities of 35 and 0.7, respectively. Conversely, there was a significant effect of salinity on the mean heart rate of H. sexdentatus (oneway ANOVA, P < 0.001; Fig. 3B). The mean heart rate of 79 ± 9 beats min −1 at a salinity of 18 (50% SW) was significantly lower than that at the two salinity extremes [0.7 (2% SW) 123 ± 8 beats min −1 ; 53 (150% SW), 131 ± 7 beats min −1 ]. There were, however, no significant differences in H. sexdentatus heart rate in any salinity relative to the full-strength SW control.
Analysis of haemolymph osmolality showed that both crabs were hyperosmotic regulators in dilute salinities ( Fig. 4A and B). In both H. crenulatus and H. sexdentatus haemolymph osmolality dropped significantly in crabs exposed for 6-h to a salinity of 18 (50% SW) relative to crabs maintained at a salinity of 35 (100% SW; overall one way ANOVA, P < 0.001 for both species). Thereafter, further dilution of the exposure medium had no further significant effect on the osmolality of H. sexdentatus haemolymph, whereas at a salinity of 0.7 (2% SW), H. crenulatus osmolality was significantly lower than that at all other salinities. In hypersaline conditions (salinity = 53; 150% SW), both species displayed a significant increase in haemolymph osmolality.
Patterns for haemolymph ions were generally similar to those for osmolality ( Fig. 4C-H). Relative to the control (salinity 35, 100% SW), crabs of both species exposed to dilute salinities displayed lower haemolymph ion concentrations, whereas crabs exposed to higher salinities displayed elevated haemolymph ion concentrations (one way ANOVAs, all P < 0.001). The threshold salinities, at which ion concentrations became significantly different, varied between ions and species. For sodium, potassium and chloride in H. crenulatus, statistically significant differences relative to the control were observed in crabs exposed to salinities of 0.7, 9 and 9, respectively (Fig. 3C, E and G). For sodium, potassium and chloride in H. sexdentatus, statistically significant differences relative to the control were observed in the haemolymph of crabs exposed to salinities of 9, 18 and 9, respectively (Fig. 4D,F and H). The single exception where an increase in exposure salinity to 53 (150% SW) did not lead to a statistically increased haemolymph ion concentration was for chloride in H. sexdentatus.
A 6-h exposure to lowered water DO had a significant overall effect on heart rate for both H. crenulatus (one-way ANOVA, P < 0.001; Fig. 5A) and H. sexdentatus (one-way ANOVA, P < 0.001; Fig. 5B). In the former species, heart rate was statistically unchanged relatively to the normoxic control (18.7 kPa), until exposure to 1.1 kPa. In crabs from this treatment, a mean heart rate of 43 ± 3 beats min −1 was observed, a value just 35% of that measured in normoxic waters (122 ± 15 beats min −1 ). For H. sexdentatus, a significant fall in heart rate with declining DO was observed at 4.7 kPa, a treatment of higher PO 2 than that observed to have the same effect in H. crenulatus. In the lowest PO 2 (1.1 kPa), the heart rate of H. sexdentatus was 33 ± 2 beats min −1 , a value 32% of that recorded in normoxic crabs of this species (103 ± 10 beats min −1 ).  A 6-h exposure to reduced PO 2 resulted in decreases in haemolymph osmolality in both crab species (Fig. 6A and B; one way ANOVAs; P < 0.001 for H. crenulatus; P = 0.002 for H. sexdentatus). All treatments displayed osmolality values that were lower than the normoxic control for both species. Haemolymph sodium concentrations also differed significantly as a function of exposure PO 2 (one way ANOVAs, both P < 0.001; Fig. 6C and D). For H. crenulatus, haemolymph sodium was maintained at normoxic control levels, until the lowest tested exposure PO 2 (1.1 kPa), where the value of 463 ± 10 mM was significantly distinct from crabs in all other waters (Fig. 6C). Conversely, in H. sexdentatus dropping the PO 2 from 18.7 to 9.3 kPa caused a significant 22% fall in haemolymph sodium (Fig. 6D). A further significant decline in H. sexdentatus haemolymph sodium was noted at 4.7 kPa. At 1.1 kPa, there was a small but significant increase, in haemolymph sodium relative to the 4.7 kPa treatment, but this value was still significantly reduced with respect to the control. Relative to normoxia, there were no significant effects of PO 2 on haemolymph potassium ( Fig. 6E and F) or haemolymph chloride (Fig. 6G and H), for either crab species. However, for H. sexdentatus, crabs from the 1.1 kPa exposure condition displayed haemolymph potassium values that were significantly reduced relative to the 4.7 and 9.3 kPa treatments.

Respiratory responses to acute salinity change
The MO 2 of H. sexdentatus acutely exposed to either hypoor hyper-saline waters was unchanged across all tested salinities (Fig. 1B). In contrast, an increase in MO 2 was observed in H. crenulatus as crabs moved from full-strength SW to 50% SW (salinity = 18; Fig. 1A). In other crab species, increases in MO 2 with declining salinity are commonly reported (Taylor et al., 1977b;Shumway, 1983;Normant and Gibowicz, 2008), although some authors have reported MO 2 increases in hypersaline waters (Ramaglia et al., 2018) or a lack of change in MO 2 with salinity (Winch and Hodgson, 2007;Theuerkauff et al., 2018). Differences between studies are likely due to diverse experimental protocols (most notably whether the measurement occurs after acute exposure or after acclimation) and differences in the osmoregulatory capacity of the test species.
However, the current findings of a transient effect or lack of response of MO 2 to salinity are distinct from previous studies in Hemigrapsus. For example, employing a protocol comparable to that of the current work, an acute (2-6 h) salinity exposure in H. takanoi resulted in a higher MO 2 in lower salinities (Shinji et al., 2009). We propose that differences in experimental outcomes reflect subtle differences in ecological niches, which are known to vary between Hemigrapsus species and between populations of the same species from different regions (e.g. McLay, 1988).
Over the course of a 6-h acute exposure neither crab species displayed changes in P crit . The P crit is widely used as an indicator of hypoxia tolerance in aquatic biota, representing the point at which an oxyregulator can no longer maintain metabolic rate and thereafter oxygen consumption declines as a function of declining PO 2 (Yeager and Ultsch, 1989). Previous studies have shown that exposure of crabs (Carcinus maenas and Carcinus aestuarii) to dilute salinities leads to an increase in P crit (Taylor et al., 1977a;Rivera-Ingraham et al., 2016). This is likely a consequence of the greater oxygen demands of crabs in lower salinities in these studies, such that the onset of the transition between aerobic and anaerobic metabolism occurs at a higher PO 2 . Given that the crabs in the current study did not show a consistent pattern of increased metabolic costs in dilute salinities, then the finding of a salinity-independent P crit is not surprising.

Cardiovascular responses to acute salinity change
Exposure of crabs to salinity change is known to induce tachycardia (Hume and Berlind, 1976;Taylor, 1977;McGaw and McMahon, 1996) species bradycardia may be observed (McGaw, 2006). However, in the current study, a change in exposure salinity had no impact on heart rate relative to controls where heart rates were monitored in full-strength SW. Tachycardia in response to salinity change has been proposed as a mechanism that facilitates oxygen uptake, thereby fuelling enhanced metabolic costs associated with osmoregulation and/or increased locomotor costs (Taylor, 1977). Consequently, our data showing an absence of tachycardia are consistent with our results showing a lack of consistent effect of salinity on MO 2 . Heart rate alone, however, does not always provide a complete picture of cardiovascular change. In crabs, cardiac output can change independently of heart rate due to altered stroke volume, such that heart rate may not accurately reflect changes in cardiovascular dynamics (McGaw and McMahon, 2003). It is therefore possible that although heart rates were relatively consistent across different salinities, cardiac output may not have been. However, even if changes in cardiovascular physiology occurred, it seems as though these acted to maintain MO 2 rather than to meet increased costs associated with hypo-or hyper-saline exposure.

Osmotic and ionic responses to acute salinity change
Both H. crenulatus and H. sexdentatus maintained haemolymph osmolality and ion concentrations below and above, those of more dilute or concentrated salinities, respectively. Although the current study only determined osmolality after 6 h (i.e. a tidal cycle) and it can take up to 48 h for osmoregulatory status to develop completely (Lovett et al., 2001), our findings were consistent with previous work on these species (Taylor and Seneviratna, 2005;Lee et al., 2010;Urzúa and Urbina, 2017).
In both crabs, haemolymph osmolality dropped significantly as animals acclimated to full-strength SW were placed in 50% SW (salinity = 18; Fig. 4A and B). The effects of salinity on haemolymph ions were more distinct and highlight H. crenulatus as the stronger ion regulator. For example, the threshold salinity at which haemolymph sodium ion concentration differed significantly from the control was 0.7 and 9, for H. crenulatus and H. sexdentatus, respectively. For potassium ion, the salinity threshold was 9 for H. crenulatus and 18 for H. sexdentatus. Distinct patterns for haemolymph osmolality and haemolymph ions within the same treatment group are likely due to changes in unmeasured osmolytes (e.g. amino acids), which are known to fluctuate with external salinity (Findley and Stickle, 1978).
The finding that H. crenulatus is the better regulator over the course of a 6-h exposure is generally consistent with previous data regarding the tolerance of these two species to salinity change. For example, Hicks (1973) noted that H. crenulatus was more tolerant to dilute salinities than H. sexdentatus at an exposure temperature of 15 • C. Conversely, Taylor and Seneviratna (2005), found that H. crenulatus was less tolerant to low salinities, although this study was conducted on early life-stages. This finding does, however, have some support from the data in the current work. At the lowest tested salinity (0.7), H. sexdentatus was able to maintain osmolality at a level statistically indistinct from that in 50% SW (18), whereas in 0.7 salinity water H. crenulatus osmolality was statistically lower than that of crabs at 50% SW. This ability to regulate in very dilute salinities has been attributed to the association of this species with freshwater inputs (Williams, 1969).
In the current work, the mass of the adult crabs differed markedly, with H. sexdentatus being three times larger than H. crenulatus. A recent study examining the capacity of H. crenulatus to regulate haemolymph sodium in response to decreasing salinity noted that larger crabs were better regulators (Urzúa and Urbina, 2017). It is therefore noteworthy that on the basis of mass differences alone, H. sexdentatus would be predicted as the better regulator. This indicates that the effects seen in the current study were not simply a consequence of body mass and that a more distinct separation of the osmoregulatory and ionoregulatory capacities of the two species may have been identified if experimental body sizes were equivalent.

Respiratory responses to hypoxia
Under conditions where crabs were exposed to declining PO 2 , resulting from their consumption of oxygen within a sealed chamber, both species initially oxyregulated. This pattern was then superseded by an oxyconforming response once P crit was reached (Fig. 2). The measured P crit values ranged from 5.5 to 7.6 kPa, in line with previous studies that have characterized P crit in intertidal crab species. For example, in Carcinus a P crit of 5.3 kPa was determined (Taylor, 1981), albeit at a slightly lower experimental temperature than that used in the current study (10 vs. 15 • C). In general, intertidal crabs display P crit values that are intermediate to the higher values measured in crustaceans that function in well-oxygenated subtidal environments (e.g. 4-11 kPa) and to the lower values determined for species that inhabit poorly oxygenated burrows (1.3-6.7 kPa; Whiteley and Taylor, 2015). Contrary to prediction, the more fossorial species, H. crenulatus, did not display a lower P crit , indicative of higher hypoxia tolerance. Given that differences in cardiovascular responses to hypoxia were discerned in the current work, this suggests that the P crit may not be a useful indicator of relative hypoxia tolerance in these species.

Cardiovascular responses to acute hypoxia
In response to declining PO 2 , both Hemigrapsus species displayed bradycardia. This is a commonly observed response to hypoxia in crabs (Bradford and Taylor, 1982;Airriess and McMahon, 1994) and is one that is usually accompanied by an increase in stroke volume (McGaw and McMahon, 2003). Together, these changes are thought to aid oxygen loading at the gills by increasing the volume of haemolymph oxygenated and its branchial residence time, whilst also facilitating oxygen unloading at the tissues (McMahon, 2001).
Although both species in the current study displayed bradycardia, the onset of this response differed. For H. crenulatus, a significant bradycardia only occurred once PO 2 reached 1.1 kPa, whereas the equivalent value for H. sexdentatus was 4.7 kPa. This suggests that H. sexdentatus is less hypoxia tolerant. In general, there is a correlation between bradycardia and hypoxia tolerance in crustaceans. In species such as Cancer pagurus, a relatively hypoxia-sensitive crab, heart rate decreases as PO 2 declines (Bradford and Taylor, 1982). However, burrowing crustaceans are highly hypoxiatolerant and often do not exhibit a bradycardia in waters of low PO 2 , instead relying on adaptations such as respiratory pigments with high oxygen affinity to maintain metabolic rate (Whiteley and Taylor, 2015).

Osmoregulatory responses to acute hypoxia
In both H. crenulatus and H. sexdentatus, exposure to reduced PO 2 resulted in decreases in haemolymph osmolality and sodium ion concentrations (Fig. 6A-D). In the freshwater prawn, a similar effect of hypoxia has been noted (Cheng et al., 2003). This was attributed to haemolymph dilution, which resulted from enhanced water influx associated with elevated ventilation rates. An alternate explanation is that the observed effects are the consequences of osmorespiratory compromise at the gill. A standard response to hypoxia in crabs is to increase ventilation rate, compensating for the reduced PO 2 by bringing larger volumes of water in contact with the gill epithelia (McGaw and McMahon, 1996). This would exacerbate diffusive exchange of ions should any small differences in extracellular and environmental ion concentrations exist. Similarly, the reduction in MO 2 as PO 2 drops means that there is reduced energy available to restore osmotic balance (Lucu and Ziegler, 2017), which could also result in changes in haemolymph osmolality and ion concentrations.
The threshold of effects of PO 2 on haemolymph osmolality was identical in H. crenulatus and H. sexdentatus. The drop in PO 2 from 18.7 to 9.3 kPa, induced significant osmolality declines in both species. However, with respect to effects on sodium ion, H. crenulatus was better able to maintain concentrations (until a PO 2 of 1.1 kPa), relative to H. sexdentatus (statistically significant drop in haemolymph sodium relative to normoxic control at 9.3 kPa). The mechanism for this is unknown, but if H. crenulatus had a superior anaerobic capacity to H. sexdentatus, then this could provide the energy to better sustain ion regulation over the duration of the shortterm 6-h exposure. These data nevertheless suggest that H. crenulatus is the more tolerant of the two species to hypoxia, consistent with cardiovascular data and its habitation of environments with greater risk of exposure to low PO 2 .

Methodological and environmental considerations
In the current study, crabs were added into exposure chambers and immediately subjected to physiological investigation. This differs from approaches where the animals are acclimated to the chambers for several hours prior to medium manipulation (e.g. Broughton et al., 2017). Consequently, the responses measured may reflect stress associated with handling. However, analysis of continuous recordings in respirometry experiments showed that any initial elevations in oxygen consumption lasted less than 20 min (data not shown). Some authors have also utilized pauses in heart rate as an indicator of reduced stress in crabs (McGaw and Nancollas, 2017). In our study, the mean time of acardia onset in control crabs was 86 min (data not shown), suggesting that crabs settled relatively quickly into chambers. Nevertheless, we cannot rule out that a component of time-integrated measurements (i.e. haemolymph osmolality and ions) could reflect handling stress.
The ability of our study to draw conclusions regarding the importance of physiology in shaping niche habitation is also limited by a lack of knowledge of the physical and chemical variability in the habitats, and more specifically the microhabitats, of the study species.

10
that the tested levels of salinity and DO are beyond the scope of those experienced by the crabs in natural settings. Consequently, interpreting physiological tolerance at physicochemical extremes in the laboratory may be misleading if environmental values for salinity and DO fluctuate over narrower ranges. Furthermore, laboratory studies involve the maintenance of animals under controlled conditions. Given that crab species exhibit physiological responses that are entrained by environmental cues (e.g. McGaw and McMahon, 1998), then this represents an additional challenge when attempting to ascribe ecological niche habitation to physiological tolerance.

Conclusion
H. crenulatus appeared to be more tolerant to low environmental PO 2 than H. sexdentatus, befitting its habitation of environments where hypoxia may occur (i.e. burrows in muddy substrates; McLay, 1988). Similarly, lower onsets of salinity effects on haemolymph sodium and potassium ion concentrations in H. crenulatus relative to H. sexdentatus hint at a greater short-term tolerance of the former species to dilute salinities, at least in the 9-18 (25-50% SW) range. This would be consistent with a crab that inhabits the lower intertidal zone, where exposure to salinity fluctuations is more commonly encountered.
Overall, however, the effects of DO and salinity on Hemigrapsus physiology were relatively minor, suggesting that factors other than physiology will contribute to the distinct niches of these two species along New Zealand coastlines. For example, the current study exposed crabs under conditions where normal behavioural responses could not be enacted. The importance of responses such as emersion and avoidance in crabs exposed to environmental stressors is well-established (McGaw et al., 1999;Bell et al., 2009) and will be of relevance in determining habitat selection in natural settings. Inter-specific interactions will also play a role. Indeed, it is notable that in Chile H. crenulatus is the sole Hemigrapsus species and occupies a wider and more exposed environmental niche than it does in New Zealand (McLay et al., 2011). It is also important to note that the current study only examined salinity and DO as stressors. Similar studies of sympatric intertidal crab species have identified temperature and tolerance to this stressor, as the major factor explaining differences in species distributions (Jensen and Armstrong, 1991;Stillman and Somero, 1996).
Finally, the current study has implications for conservation of these species. The coastal habitats of Hemigrapsus in the Canterbury region of New Zealand are subjected to anthropogenic change. For example, the Avon-Heathcote Estuary/Ihutai receives nutrient inputs that result in water quality metrics that exceed regulatory trigger values and which lead to extensive anoxia (Bolton-Ritchie and Main, 2005). Beach habitats are exposed to coastal erosion, which will worsen with predicted sea level rise (MfE, 2017). This can result in changes in tidal profiles and narrowing of ecological niches. This is exacerbated by modifications of coastal infrastructure such as sea walls, which may further condense optimal habitat and expose animals to more severe fluctuations in environment (Gittman et al., 2016). The current work indicates that the Hemigrapsus crab species found on New Zealand coasts have sufficient physiological plasticity to enable them to adjust niche in response to environmental change. However, shifts in ecological niches may increase inter-specific competition and ultimately could lead to exclusion of the less robust species.