Introduction

The Eastern oyster, Crassostrea virginica, has a remarkable latitudinal distribution range of ~4,000 km in the Western Atlantic (Carriker and Gaffney 1996). The southern limit is in the Gulf of Mexico (27°N), and the range extends northward to the Gulf of St. Lawrence (48°N), Canada (Fig. 1). Accordingly, the species is highly tolerant of extreme ambient temperatures; when exposed at low tide, it can survive temperatures up to 49 °C (Galtsoff 1964) as well as freezing temperatures (Loosanoff 1965).

Fig. 1
figure 1

Map showing three locations of Crassostrea virginica referred to in text. Water temperature time series provided for Gulf of St. Lawrence estuaries (DFO unpubl. data, 2000–2001 daily means), Long Island Sound (NOAA National Data Buoy 44039, 2005–2006 monthly means) and Gulf of Mexico (NOAA National Data Buoy 42020, 2005–2006 monthly means)

Full size image

The ability to survive such extreme temperatures implies physiological adaptation. While several studies have examined the effect of low temperature on the feeding behaviour of bivalves, only a few rather dated studies have dealt specifically with C. virginica. Oysters in these experiments were from the east coast of the United States. Following the monitoring of stomach contents of Edge Cove (40°N) oysters, Nelson (1921) reported that food ingestion was negligible below a critical temperature of 5.6–7.2 °C. This finding was supported by Galtsoff (1926, 1928a) who reported no pumping activity below 5.0–7.6 °C for oysters originating from Chesapeake Bay (37°N), Long Island Sound (41°N) and Wellfleet Harbour (42°N). Loosanoff (1958) subsequently found that a small number of individuals from Long Island Sound (41°N) pumped water and produced faeces at lower temperatures. He concluded that the “feeding of oysters below 5 °C occurs only as an exception.” Since the publication of these landmark papers, 5 °C has been widely accepted as the lower feeding threshold for C. virginica.

The applicability of the above findings to natural populations warrants caution for a number of reasons. For instance, the conclusion offered by Loosanoff (1958) was based on short-term (24-h) exposures, and the author himself cautioned that his results might “not necessarily be representative of longer exposures.” A second issue is that the diet was unnaturally supplemented with Chlorella sp., whereas subsequent studies showed that ecologically relevant behaviour in bivalves is best observed when they are offered natural seston (Hawkins et al. 1998; Cranford et al. 2005). Finally, methodological details aside, the pioneering work on C. virginica may not apply to populations residing throughout its wide range. Oysters in the above experiments were from the east coast of the United States (37–42°N), where water temperature is seldom <5 °C (Fig. 1, inset). In contrast, at the northernmost range limit of the species (48°N), estuarine temperatures remain stable within a narrow range of −1.5 to 0 °C over five consecutive months (November–April). Conceivably, such acute cold conditions may have led to the development of a northern strain of C. virginica, characterised by a very low temperature threshold for feeding. Such adaptability would be consistent with the latitudinal-compensation theory, which predicts that individuals at higher latitudes compensate for the temperature-associated slowing of metabolic processes by increasing physiological functions (Dittman 1997; Pörtner 2002; Kokita 2004; Sukhotin et al. 2006). For bivalve suspension feeders, latitudinal compensation may be expressed as the ability to continue feeding at very low temperatures (Petersen et al. 2003).

The present study reports on the winter–spring gaping behaviour of C. virginica exposed to natural conditions at its northernmost distribution limit (48°N). Valve gaping signals the activation of a complex nervous mechanism involving the heart and adductor muscles (Taylor 1976), resulting in the bivalve exposing itself to the ambient environment and exercising metabolically demanding processes, such as the collection and assimilation of food particles. Given that the oysters were from the northernmost population in the Gulf of St. Lawrence, we hypothesised that their valves would open at lower temperatures than those previously reported for southern populations.

Materials and methods

Study site

St. Simon Bay (47° 44′ N, 64° 47′ W) is an oyster-producing coastal inlet in the province of New Brunswick (Electronic Supplementary Material Appendix 1). Approximately 18 % of its surface area is allocated to oyster farming (Comeau et al. 2006). Off-bottom culture, in which oysters are held inside floating PVC bags or floating cages, is the predominant farming technique. In winter, floaters are removed to lower stocks onto the seabed where they are protected from the thick (~1 m) ice cover, which generally sets in early January and melts in April. At other times, however, stocks are suspended near the surface in a relatively warm and phytoplankton-abundant environment, thereby enhancing shell growth and shortening the production cycle.

The oysters used in this study originated from a natural set in Caraquet Bay (Electronic Supplementary Material Appendix 1), a neighbouring embayment where they grew to a SL of ~15 mm before being transferred to St. Simon Bay. In February 2010 and 2011, bags of market-size oysters (SL > 64 mm) were removed from the St. Simon seabed through holes in the ice. The oysters were transferred to the Étang Ruisseau Bar Ltd. (ERB) wet holding facility on the shores of St. Simon Bay. Oysters were housed in two holding tanks (~150 l) that were continuously supplied with unfiltered seawater pumped from the main channel entering St. Simon South.

Environment

Inside the windowless ERB facility, artificial lighting was generally limited to a 9-h period from 8:00–17:00 h (5 days week−1). Occasionally this background lighting was supplemented by natural light entering the building through the large receiving doors. While lighting intensity was not rigorously controlled, it was continuously monitored using Hobo UA-002 light loggers (Onset Computer Corporation, Massachusetts, USA), which were placed 2 cm above the holding tanks.

Water temperature was monitored using Minilog-TR loggers (Vemco, Nova Scotia, Canada). Salinity was determined from electrical conductivity measurements taken using a field probe (YSI 6560, Ohio, USA). Relative fluorescence was measured every 30 min using a YSI 6025 (YSI, Ohio, USA) chlorophyll sensor in 2010, and a Phytoflash active fluorometer (Turner Design, California, USA) in 2011. Relative fluorescence units (RFU) were corrected for instrument drift, and converted into chl a concentrations (μg l−1) using the equation 7.2416 × RFU−1.4095 (r 2 = 0.84, P < 0.05) for the YSI fluorometer, and the equation 12.593 × RFU−1.2057 (r 2 = 0.74, P < 0.05) for the Phytoflash fluorometer. These equations were derived from a series of chl a extractions during the study. In brief, duplicate water samples (100 ml) were collected bi-weekly from the pipe supplying the holding tanks. Chl a was filtered onto GF/C glass microfiber filters, extracted in 90 % acetone and thereafter measured fluorometrically (Parsons et al. 1984) using a Turner Aquafluor handheld fluorometer (Turner Designs, Sunnyvale, California, USA).

Gaping

Sixteen (2010) and 14 (2011) oysters ranging in SL from 63 to 85 mm were randomly selected to monitor gaping activity. Following Nagai et al. (2006), a coated Hall element sensor (HW-300a, Asahi Kasei, Japan) was glued to the left valve at the maximum distance from the hinge. Then, a small magnet (4.8 mm diameter × 0.8 mm height) was glued to the right valve, directly above the Hall sensor (Electronic Supplementary Material Appendix 2). After the glue had dried, the wired oysters were placed on their left valve in the holding tanks, and left to acclimate for 7 days prior to the onset of measurements. The magnet and the Hall element weigh 0.1 and 0.5 g, respectively. For comparison purposes, a small (6 mm diameter) live barnacle weighs approximately 0.12 g.

Gaping was continuously recorded at a frequency of 10 Hz for 113 days in 2010 (20 February–12 June), and for 34 days in 2011 (1 April–5 May). The magnetic field or flux density between the sensor and magnet was a function of the gap between the two valves. The magnetic field in the form of output voltage (μV) was acquired by strain recording devices (DC 104R, Tokyo Sokki Kenkyujo Co., Japan). Output voltage was converted into valve opening by applying conversion algorithms specific to each sensor assembly. At the end of the monitoring phase, the adductor muscle was severed, and small calibration wedges were manoeuvred between the two valves at the point farthest from the hinge. Wedge height was 1–6 mm. The relationships between voltage and wedge height (i.e. valve opening) were non-linear and strong (r 2 > 0.92). Valve opening (mm) data were then converted into gape angles (θ in degrees) using the following equation (Wilson et al. 2005):

$$ \theta = 2\arcsin \left( {\frac{0.5\,W}{L}} \right) \times 100 $$

where W is the valve opening (mm) and L (mm) is the bivalve’s shell length.

Statistics

The 10-Hz sampling rate generated a large dataset, containing ~97 × 106 gaping measurements for each of the 16 oysters monitored in 2010, and ~29 × 106 measurements for each of the 14 oysters monitored in 2011. Taking into account the amount of processing time and the paper’s focus on seasonality, we explored whether data thinning had any impact on the key metrics. Exploratory analyses were performed on time series of varying length (1–30 days). Lowering the sampling interval from 10 Hz down to one measurement min−1 caused no significant effect on the population means, and a dataset abridged in this way was used.

The relationship between temperature and gape angle was investigated using non-autocorrelated data. Temperatures were rounded to their nearest degree and, for each degree, the maximum gape angle of each individual was retained for analysis, irrespective of the time the measurement was taken. The relationship was best described using a four-parameter logistic model: \( y = A + \left( {B - A} \right)/1 + e^{{\left( {\left( {{\rm xmid} - {\rm x}}\right)/{\rm scal}} \right)}} \), where y and x represent the maximum gape angle and temperature, respectively; A is the lower asymptote; B is the upper asymptote; xmid is the inflection point (i.e. value of x where y = [A + B] / 2); and scal is a scale parameter on the x axis or, more specifically, the distance on the x axis between the inflection point and the point where y ≈ 0.75 × B. An analysis of the residuals in this non-linear model indicated large variation amongst individuals and particularly, an increasing variation in the residuals as the temperature and the fitted values increased. Random factors were added to the model in order to address the variation amongst individuals and patterns in the residuals. Therefore, the model became a non-linear mixed effects model, designed to estimate not only the average behaviour in the population, but also the variability amongst and within individuals. To determine which of the four parameters required a random factor, a logistic model was built for each individual and parameter variability amongst individuals was evaluated by plotting 95 % confidence intervals (Pinheiro and Bates 2000, 2009). A random factor was judged necessary for parameters with confidence intervals that rarely overlapped from one individual to another. It was determined that three of the four parameters required a random factor: B, xmid and scal. Finally, the effect of year (2010 vs. 2011) was tested by comparing models with and without year as a factor, using the likelihood ratio test, which follows approximately a chi-square distribution. All analyses were conducted in R (R Development Core Team 2012) using the nls and nlme functions of the nlme package (Pinheiro et al. 2012).

Results

Marked behavioural changes at the individual level were detected over time. Three distinct phases were apparent when examining the gape angle time series (Fig. 2a): quiescent (winter), awakening (spring) and active (spring). Monitoring was initiated during the quiescent phase in both study years. Recorders captured 42 days of this phase in 2010, and 21 days in 2011. Quiescence was defined as the period when valves were either closed or slightly opened (<2°).

Fig. 2
figure 2

Gape angle of a 74-mm Crassostrea virginica a over a 113-day period (20 February 2010–12 June 2010) and b over a 5-h period. Highlighted phases (quiescent, awakening and active) were clearly discernible for all oysters in study

Full size image

In contrast, valves abruptly opened towards maximum angles during the awakening phase (Fig. 2b). These sudden shifts in gaping angles started on April 3 in 2010 and April 22 in 2011. Moreover, there was a striking level of synchrony amongst individuals, with group awakenings being apparent (Table 1). For example, following the 42-day quiescent period in 2010, 50 % (8/16) of the oysters awoke in just 6.6 h. Similarly, in 2011, 50 % (7/14) of the oysters awoke within a 6.6-h period. All monitored oysters awoke over a period of 5.7 days in 2010, and 4.1 days in 2011.

Table 1 Timing of spring awakening for each oyster
Full size table

There was no apparent link between oyster awakening and the melting of winter ice or development of phytoplankton blooms. The salinity data indicated that the melting of winter ice started approximately 10 days before awakening (Fig. 3a). Similarly, there was no indication that food was limiting during quiescence or that awakening was a response to increasing food resources. No significant correlation was found between chl a concentration and gape angle. During quiescence, chl a concentrations averaged 3.7 μg l−1 (SE = 0.2, 2010) and 4.5 μg l−1 (SE = 0.1, 2011); a few chl a peaks were recorded during the period of awakening in 2010; however, these peaks were absent in 2011 (Fig. 3b, c).

Fig. 3
figure 3

a Salinity over a 26-day period in 2011 (salinity recorder malfunctioned in 2010, data not shown). Chl a series b over a 12-day period in 2010 and c over an 11-day period in 2011. Temperature series d over a 12-day period in 2010 and e over an 11-day period in 2011. Dotted lines delimit awakening periods

Full size image

In both study years, the awakenings occurred during warming periods of a similar magnitude (Fig. 3d, e). A closer examination indicated that the awakenings occurred during or shortly after ebb tides, when the flushing of cold oceanic water from the embayment resulted in periodic increases in temperature. However, the characteristics of the temperature increases were variable, ranging from 0.9 to 4.2 °C (mean = 2.5, SE = 0.2) and in rate from 0.18 to 0.66 °C h−1 (mean = 0.41, SE = 0.3). Similarly, absolute temperature at the onset of awakening varied by a few degrees amongst individuals (Fig. 4). While no single temperature can be linked to all awakenings, an influence of temperature is suggested by the discovery of a significant relationship between temperature and the maximum gape angle. This relationship was best described using a four-parameter logistic model. All individuals (30 oysters) were pooled in the analysis given that the model output was similar for both years (χ2 = 7.2, P = 0.12). Model results are shown in Table 2 and Fig. 5. The lower asymptote A, which represents the maximum gape angle during the quiescent phase, is significantly different from 0, and averages 0.49° with very low between-individual variability. At the opposite end of the curve, the upper asymptote parameter B was also significantly different from 0, averaging 5.88°. However, between-individual variance was quite elevated for B (σ 2 B  = 1.562). On average, the inflection point xmid was reached when the temperature was 3.39 °C, with a between-individual variance of 1 °C (σ 2xmid  = 1.042). Based on individual trend lines leaving their lower asymptotes, it could be said that quiescence occurred at temperatures <0.2 °C and that awakenings began at temperatures of 0.2–4.0 °C (mean = 2.2, SE = 0.2). Similarly, trend lines reaching their upper asymptotes suggest that awakenings were completed at temperatures of 2.8–6.6 °C (mean = 4.8, SE = 0.2).

Fig. 4
figure 4

Maximum gape angle as a function of temperature. Plots represent single oysters in 2010 (top panel) and 2011 (bottom panel)

Full size image
Table 2 Parameter estimates for maximum gape angle and temperature in non-linear mixed effects model
Full size table
Fig. 5
figure 5

Maximum gape angle as a function of water temperature. Logistic trend lines shown for each of 30 oysters monitored in 2010 and 2011. Population trend (bold line) defined by equation \( y = {{0.49 + \left( {5.88 - 0.49} \right)} \mathord{\left/ {\vphantom {{0.49 + \left( {5.88 - 0.49} \right)} {1 + e^{{\left( {{{\left( {3.39 - x} \right)} \mathord{\left/ {\vphantom {{\left( {3.39 - x} \right)} {0.29}}} \right. \kern-\nulldelimiterspace} {0.29}}} \right)}} }}} \right. \kern-\nulldelimiterspace} {1 + e^{{\left( {{{\left( {3.39 - x} \right)} \mathord{\left/ {\vphantom {{\left( {3.39 - x} \right)} {0.29}}} \right. \kern-\nulldelimiterspace} {0.29}}} \right)}} }} \), where y and x represent maximum gape angle and temperature, respectively

Full size image

Once awakened, the oysters remained open most of the time. The study captured 65 days of this active phase in 2010, but only 8 days in 2011. During this phase, oysters remained open an average of 68.6 % (2010, SE = 3.2, n = 16) and 79.7 % (2011, SE = 5.5, n = 14) of the time. The durations of openings and closures were highly variable, both temporally and amongst individuals. For instance, with regard to the longest series available (65 days in 2010), the duration of opening ranged from 1 min to 13.8 days, while the duration of valve closure ranged from 1 min to 3.7 days. Figure 6 shows the average duration of opening and closure for each individual oyster. There was a tendency for individuals to maintain longer periods of opening than closure (paired t test applied to sample means, P < 0.001).

Fig. 6
figure 6

Average duration of valve openness and closure (SE) for each oyster monitored during active phase of 2010 (16 April–12 June)

Full size image

It was noteworthy that gaping followed a circadian rhythm, exhibiting a 24-h oscillation period (Fig. 7). It appears the rhythm was synchronised with shifting light conditions inside the holding facility. Both the percentage of time that valves were opened and gape angle progressively increased during the daytime hours, and peaked at the onset of darkness.

Fig. 7
figure 7

a Light intensity and b gaping activity as a function of time of day during active phase of 2010 (16 April–12 June). Data points represent sample averages (SE) (n = 16). Shaded areas indicate periods of darkness

Full size image

Discussion

Winter quiescence

We initially postulated that the feeding apparatus of Crassostrea virginica has adapted to the extreme cold conditions of eastern Canada, at the northernmost distribution limit of the species, but no evidence was found to support that hypothesis. During winter (<2 °C), the studied oysters either closed themselves off completely from the surrounding environment or slightly opened their valves, with maximum gape angles averaging only 0.49°. To date, winter quiescence or dormancy in cold temperate areas has been described for suspension feeders such as bryozoans, hydrozoans, ascidians and holothurians (Costelloe and Keegan 1984; Singh et al. 1999; Coma et al. 2000). For bivalves, however, it seems that detailed descriptions of this phenomenon are largely restricted to populations inhabiting polar regions (Brockington 2001; Rodrigues et al. 2007). The complete valve closures recorded in our study are consistent with biochemical pathways that can sustain energy production anaerobically. Like other bivalve species, C. virginica can rely on the coupled fermentation of glycogen and aspartate, with succinate and alanine accumulating as end products (Gade 1983; Stickle et al. 1989; Greenway and Storey 1999; Bumett and Stickle 2001). The small valve openings in turn may signal the flushing of dissolved waste products, and also the onset of aerobic metabolism. Total lipids in the digestive glands of the cultivated oysters used in our study fell by a factor of approximately two over the course of the winter (E. Mayrand personal communications). The narrow valve openings probably allow sufficient gas exchange to facilitate the metabolism of lipids. Therefore, it appears that quiescent oysters do require a supply of oxygen, though the demand may be low because of metabolic suppression.

It is doubtful that the slight valve openings during quiescence signalled active feeding. Under laboratory settings, only 5 % of oysters cleared phytoplankton cells in winter-like conditions (Loosanoff 1958; Pernet et al. 2007; Comeau et al. 2008). It also seems unlikely that the small openings were indicative of deteriorating health conditions. Wang and Amiro (1977) visually monitored the valve openings of oysters maintained in running natural seawater at 1–2 °C over a six-month period during winter. They reported that oysters rarely opened their valves, and those that opened their valves frequently were more likely to die. In the current study, however, none of the experimental oysters died. We studied cultivated oysters that were well conditioned prior to the onset of winter because they were grown in the upper water column, where food fluxes are substantially higher than fluxes near the sea bed (Comeau et al. 2010). In comparison, wild and presumably poorly conditioned oysters originating from the Prince Edward Island sea bed (47°N) were used in the study by Wang and Amiro (1977).

Limited food resources (Barnes and Clarke 1995; Coma 1998; Ribes 1999) and low temperatures (Clarke 1993) have been suggested as being the main factors that trigger winter dormancy in other suspension feeders. The influence of food seems unlikely for C. virginica, because chl a levels were quite elevated (~4 μg l−1) in the winter and comparable to those observed in summer. Moreover, there is laboratory evidence that C. virginica enters quiescence despite being fed rich laboratory diets (Loosanoff 1958; Pernet et al. 2007). Conversely, it is possible that quiescence was induced by low temperatures reducing the rate of chemical reactions (Clarke 1993; Clarke 1998) and/or increasing the viscosity of the water (Podolsky 1994). This interpretation is supported by the ciliated gill cells of C. virginica having a limited ability to transport particles or produce a current at low temperatures (Galtsoff 1928b; Dittman 1997), in addition to the positive correlations between gaping and temperature found in the present study.

Based on an archival temperature series (DFO, unpubl. data), we estimate that the quiescent phase may last as long as 160 days. A lack of grazing over such an extended period of time has implications for oyster productivity. For instance, it may be a major explanatory factor for the slow growth of individuals, that is, it takes a Gulf of St. Lawrence oyster 4–7 years to reach the legal market SL of 76 mm, whereas oysters in the warm waters of the Gulf of Mexico may reach the same size in just 2 years (Medcof 1961; Galtsoff 1964). The capacity of oysters to adapt to starvation through anaerobiosis and metabolic suppression undoubtedly favours survival during quiescence. In the case of starved Pacific oysters, Crassostrea gigas, storage reserves are reportedly depleted after 113 days (Ren and Schiel 2008), and cumulative mortalities reach 30 % after 405 days (Whyte et al. 1990). Death is linked to the utilisation of structural tissue (Kooijman 2000). Oysters in those experiments were maintained at 10–18 °C. Presumably, there is a slower decline of storage energy in C. virginica at its northernmost limit, where temperatures in winter remain <0 °C.

Spring awakening

One of our most striking findings was the ability of oysters to quickly change from a quiescent to an active state. The process of awakening may be described as a brief transitional phase, when the oyster abandons its quiescent state and gradually opens its valves, attaining a fully open state within hours. Awakening may be viewed as an immediate response to favourable conditions in which oysters may fully resume aerobic metabolism, ventilate their gills and collect food particles. Our study provided some interesting insights about the awakening phase, notably with respect to the synchrony amongst individuals. Following a long quiescent period, many individuals awoke only minutes apart from one another. This observation indicates that a common cue might be present. Although speculative, waterborne chemical(s) may have played a role, perhaps prompting the group awakenings, much in the same manner that pheromone peptides stimulate females to release their eggs in mass spawning or that various metabolites attract settling larvae (Zimmer-Faust and Tamburri 1994; Walch et al. 1999; Hadfield and Paul 2001; Bernay et al. 2006). At present, the available data suggest that temperature is a main factor controlling valve movement and therefore presumably also the awakenings. This interpretation is consistent with a reported causal relationship between temperature (range 0–38 °C) and the proportion of open oysters in experimental groups (Loosanoff 1958). Our logistic models predict an onset of awakening at a temperature of 0.2–4.0 °C (mean = 2.2, SE = 0.2). Other variables are almost certainly at play given the variation in the awakening temperature amongst individuals. The observed variance may be attributed to intraspecific variations in biochemical (membrane lipids) and physiological (basal metabolic rate) adaptations to temperature (Pernet et al. 2008). Oysters exerted their maximal gaping angle as soon as temperature reached 2.8–6.6 °C (mean = 4.8, SE = 0.2). Regardless of the exact cue to abandon the quiescent state, the reported observations are novel and may serve fishery management. For instance, they could be useful for the management of shellfish closures during toxic algal blooms in the spring (Bates et al. 2002). The information could also serve the aquaculture industry. In autumn, cultivated oyster stocks are lowered onto the sea bed where they are protected from the thick ice, but nonetheless exposed to the deleterious effects of siltation, and potentially become buried in the soft mud. This environment probably causes little stress to the oysters as long as they remain quiescent with their valves closed. However, at the onset of the spring awakening, exposure to a silty environment could clog the gill apparatus, potentially proving lethal within 3 (Hsiao 1950) to 11 days (Comeau et al. 2011). Consequently, it would be advisable to re-suspend stocks in the upper water column at the onset of the awakening period.

Post-awakening (active phase)

In the post-awakening period, water temperature no longer exerted an influence on gaping. Valves remained open 68.6 % of the time (over a 65-day period in 2010). In comparison, other investigators reported that valves were open for 35.1 % (Higgins 1980), 40–60 % (Brown 1954), 71.3 % (Galtsoff 1928b) and 94.3 % (Loosanoff and Nomejko 1946; Higgins 1980) of the time. The lowest and highest percentages relate to unfed and continuously fed (algal monoculture) oysters, respectively, whereas our result (68.6 %) is based on naturally fed oysters. The reason oysters intermittently closed their valves and isolated themselves from the environment is unclear. Closures lasted as long as 3.7 days, and we found no indication that they were associated with depleted phytoplankton resources. It is possible that closure was induced by low food quality and/or digestion processes, consistent with the view that feeding is subject to physiological regulation in accordance with food resources and nutritional needs (Morton 1973; Bayne 1998; Cranford 2001).

Finally, an unexpected result of this study was the observation of a diurnal rhythm in valve openness, with the greatest gape near the end of the afternoon and least in the early morning. The existence of such a rhythm has been long debated without consensus (Orton 1929; Loosanoff and Nomejko 1946; Brown 1954; Higgins 1980). In our study, we noted that the rhythm was synchronised with the light-dark regime inside the holding facility, although light intensity was not rigorously controlled. It is well documented that the light cycle is a common extrinsic driver or synchroniser (zeitgeber) that regulates diurnal rhythms in many animals, including bivalves (Braun and Job 1965; Salánki 1966; McCorkle et al. 1979; Englund and Heino 1994; Wilson et al. 2005; Robson et al. 2010; Schwartzmann et al. 2011). The larvae of C. virginica have a pair of light-sensitive eyespots that gradually degenerate after metamorphosis (Baker and Mann 1994); however, this species might retain a light sensitivity via photoreceptive cells in its mantle tissue (Kennedy 1960; Morton 2001).

In conclusion, no behavioural evidence was found to support the hypothesis that the feeding apparatus of C. virginica has adapted to the extreme cold conditions of eastern Canada, which represents the northernmost range limit of the species. The study, however, provided new insights into the quiescent, awakening and active phases of C. virginica, which could have potentially beneficial applications for shellfish management practices.