Assessment of swimming behavior of the Pacific oyster D-larvae (Crassostrea gigas) following exposure to model pollutants


This study describes an image analysis method that has been used to analyze the swimming behavior of native oyster D-larvae (Crassostrea gigas) from the Arcachon Bay (SW, France). In a second time, this study evaluated the impact of copper and S-metolachlor pollutants on D-larvae swimming activity and the possible relationship between developmental malformations and abnormal swimming behavior. Analyses in wild and cultivated oyster D-larvae were investigated during two breeding-seasons (2014 and 2015) at different sampling sites and dates. In controlled conditions, the average speed of larvae was 144 μm s−1 and the maximum speed was 297 μm s−1 while the trajectory is mainly rectilinear. In the presence of environmental concentration of copper or S-metolachlor, no significant difference in maximum or average larval speed was observed compared to the control condition but the percentage of circular trajectory increased significantly while the rectilinear swimming larvae significantly declined. The current study demonstrates that rectilinear trajectories are positively correlated to normal larvae while larvae with shell anomalies are positively correlated to circular trajectories. This abnormal behavior could affect the survival and spread of larvae, and consequently, the recruitment and colonization of new habitats.


Behavioral analysis is increasingly used to study the effects of chemicals and drugs on humans and other mammals. In recent years, many biological early warning systems have been developed that evaluate the behavioral responses of aquatic organisms to water quality (Melvin and Wilson 2013; Garaventa et al. 2010; Van der Schalie et al. 2001). In contrast, effects of contaminants on aquatic invertebrate behavior are less frequently studied when compared to developmental or reproductive toxicology (Melvin and Wilson 2013; Scott and Sloman 2004). However, behavioral indicators of toxicity would appear ideal for assessing the effects of pollutants on aquatic organisms since they link physiological function with ecological processes (Scott and Sloman 2004). In aquatic toxicology, the link between behavioral science and impacts of toxic substances has only taken on a real importance in the last decade (review in Faimali et al. 2017). Behavioral responses are the direct result of adaptations to environmental variables. A selective response is permanently adapted by direct interaction with the physiological aspects of the chemical and physical social environment. In recent years, much progress has been made in developing technological tools available for quantifying behavior (Lv et al. 2013), and the study of behavioral parameters is now a valuable tool to identify and assess the effects of exposure to environmental stressors. Many studies have focused on the behavioral study of fish larvae or adults (Handy et al. 1999; Kazlauskiene et al. 2010; Caldwell et al. 2013; Le Bihanic et al. 2015; Sommers et al. 2016; McCallum et al. 2017; Martin et al. 2017). Very few studies have focused on mollusks and more particularly on bivalves (review in Faimali et al. 2017). However, bivalve mollusks such as mussels, clams, and oysters contribute significantly to global aquaculture production, and the same species have been used as sentinels to monitor pollution in coastal marine waters. The Pacific oyster Crassostrea gigas (Thunberg) is ranked number one in the world in terms of aquaculture production, with an estimated 625,925 tons produced in 2014 (FAO 2019). Despite living attached for most of their life stages, bivalves have a swimming larval stage, which plays an important part in their colonization of the environment. In C. gigas embryos incubated at 20 °C, the first movements were observed in trochophore at 6.5 h post-fertilization (hpf) and most of individuals (85%) can swim at 11.5 hpf (Suquet et al. 2013). Up to now, only a few works has focused on the swimming behavior of the early life stage of marine bivalves (Mileikovsky 1973; Troost et al. 2008) and only a few ones have investigated the abiotic factors as pH, salinity (Suquet et al. 2012, 2013) and pollutants (Horiguchi et al. 1998) on swimming capacities of bivalves. To our knowledge, none investigated the impact of pollutants on oyster larval swimming capacities.

To analyze the possible impact of pollutants on oyster larval behavior, copper and metolachlor, both pesticides appearing frequently in marine coastal water, were tested. Copper (Cu) is among the most hazardous metals for bivalve larvae (His et al. 1999). Although Cu is an essential micronutrient for living organisms (Festa and Thiele 2011), it can be toxic above a certain concentration depending on the organism (Flemming and Trevors 1989). All over the world, Cu is used both as fungicide in fruit culture and as part of antifouling paint to prevent aquatic organisms from attaching themselves to the hulls of vessels (Turner 2010). Copper concentrations in the Arcachon Bay have been increasing for several years and current levels exceed micrograms per liter (Gamain et al. 2016). The impact of copper on aquatic organism behavior has previously been documented on fish (Sommers et al. 2016). Pesticide contamination in Arcachon Bay is made up predominantly of metolachlor herbicide (approximately 10 ng L−1) and its metabolites (Auby et al. 2007; Tapie et al. 2017; Gamain et al. 2016). Arcachon Bay is the end recipient of several rivers draining a watershed of 4138 km2 (Auby et al. 2014), dominated by agriculture and urban areas. Metolachlor is used in agriculture to control pre-emergent and early post-emergent broadleaf and grass weeds and is regularly detected in surface and groundwater.

The main objective of this paper is to study the behavior of D-larvae of oysters (C. gigas) in response to copper or S-metolachlor exposure. To do this, a method based on image analysis was developed. The open-source program ImageJ ( and the ImageJ Plugin wrMTrck were used. This plugin was initially developed to study the mobility of the nematoda Caenorhabditis elegans while swimming in liquids (Nussbaum-krammer et al. 2015), and presents an interesting similarity with swimming larvae. The method allows analysis of multiple parameters, including total traveled length, distance, and speed of larvae.

This paper aims to (1) describe an image analysis method that has been used to analyze D-larvae behavior, (2) apply this process to assess the impact of Cu and S-metolachlor pollutants on D-larvae behavior, and (3) investigate possible relationship between developmental malformations and abnormal swimming behavior.

Materials and methods

Method based on imagery

Image acquisition

To analyze the behavior of the native oyster, wild and cultivated oysters were collected during the oyster breeding period (July and August 2014 and 2015) from three different sites in the Arcachon Bay: Comprian, Les Jacquets, and Banc d’Arguin (Fig. 1). Comprian is located close to the River L’Eyre and therefore is under the influence of fresh water inputs in contrast to Les Jacquets or to Arguin. The reference water was collected near the Banc d’Arguin. Adult oysters were brought back to the laboratory in tanks filled with seawater from the sampled site and were kept overnight in this water, which was continuously aerated and kept at 12 °C.

Fig. 1

Sampling site for mature oysters in the Arcachon Bay (SX France). 1 = Les Jacquets, 2 = Comprian, and 3 = Banc d’Arguin

Mature oysters (male and female) were induced to spawn through thermal stimulation, alternating immersion in filtered seawater (FSW) of 15 °C and 28 °C for 30 min. Spawning males and females were individually isolated in beakers containing 500 mL of 0.2 μm FSW at spawning temperature. They were left undisturbed for 15 min and then removed from beakers. Eggs and sperm from two individuals were selected to give a single pairing. Sperms and eggs were sieved separately through 50-μm and 100-μm meshes (Sefar Nitex), respectively, to eliminate debris and feces. Sperm mobility was checked and the number of eggs was counted under a microscope (LEICA DME) at a magnification of 100. Eggs were fertilized with sperm at a ratio of 1:10 (egg/sperm), homogenized with an agitator to prevent polyspermy. Fertilization success was verified under a microscope, and embryos were then counted and transferred to an 24-well microplate (Greiner Bio-One, Cellstar free of detectable DNase, RNase, human DNA, and pyrogens) for embryotoxicity assays. The embryotoxicity assay has been described in details by His et al. (1999) and normalized (AFNOR 2009). Fertilized eggs (around 300 eggs) were exposed in wells containing 2 mL of toxicant solution. The eggs density was slightly changed compare to the recommended AFNOR document specifying densities between 20,000 and 50,000 embryos per liter. These microplates were incubated in a climatic chamber at 24 °C for 24 h in the dark.

After 24-h incubation, a 2-min film in the light was recorded by means of a microscope (Nikon Inverted Microscope Eclipse TS 100/TS100-F, TS100LED MV/MV F-TS100 LED) and acquisition software (NIS Element D). Recordings were in MPEG format at 100 frames per second. The swimming behavior of about 400 control larvae was analyzed in order to determine their average speed and their maximum swimming speed, their immobility threshold as well as their typical swimming trajectory.

Description of the image analysis process

A freeware (VirtualDub) for video conversion was used to subsample the film to 4 fps and convert it to AVI format. The AVI format was used within ImageJ. The entire flow of analysis as well as the expected results are schematized in Fig. 2. The films (.avi) were opened as a stack of images (Fig. 2a) and converted into grayscale (Fig. 2b). The entire stack of images was then converted into a binary stack of images using the maximum entropy (Kapur et al. 1985) of their histogram to establish a threshold (Fig. 2c). When running the wrMTrck plugin, proper settings to track oyster larvae were selected (Table 1).

Fig. 2

Image analysis. a Original image. b Grayscale image. c “Binarized” image. d Traces of larval path. e Quantified parameters

Table 1 Optimized settings for wrMTrck plugin

As a result, each tracked larvae was assigned a number, with each number used to identify the tracked larvae in the result file. The result file (Fig. 2e) included: tracking number, sum of all movement vectors (pixel), distance between the start and the finish position (pixel), number of frames, first frame, total time (seconds), maximum speed (pixel/s), mean and standard deviation of the surface of a larva during the tracked period (pixel2), mean and standard deviation of the perimeter of a larva during the tracked period (pixel), average speed, body length per second, average X position of the larva during the tracked period, and average Y position of the larva during the tracked period. A diagram showing all tracked larval paths detected throughout the video (Fig. 2d) was also generated.

Using this diagram, it was possible to identify three different types of larval path: (1) Rectilinear, defined as a trajectory in which length and distance are fairly similar (track 1 Fig. 3); (2) circular, defined as a trajectory in which length is significantly greater than distance (track 2 Fig. 3); and (3) motionless, in which length is relatively short (track 3 from Fig. 3). It is important to note that a single larva may be detected multiple times, since they can exit and enter the field of view (track 4 from Fig. 3). Furthermore, two larvae may collide; if it is the case, their larval path may be altered (track 5 from Fig. 3).

Fig. 3

Gathering drawing with different traces of larval paths: 1 = rectilinear, 2 = circular, 3 = motionless, 4 = larval enters and exits field of view, 5 = collision between two larvae

Developmental abnormalities

Following 24 h of incubation, 25 μL of 1% buffered formalin were added to each well and the percentage of abnormal larvae was recorded. One hundred individuals per well were directly observed under an inverted microscope (Nikon eclipse TS100/TS100-F; TS100 LED MV/TS100 LED-F MV) to determine the number of abnormal D-shell larvae according to the criteria described in His et al. (1999) and AFNOR (2009). An important prerequisite for this test is the presence, in control conditions (24 °C in the absence of contamination) of less than 20% of abnormal larvae. In this experiment, four different couples were used and four replicates were performed for each condition.

Chemical analyses

Reference toxicants (CuSO4 and S-metolachlor) and formalin were purchased from Sigma-Aldrich Chemical (St. Quentin Fallavier, France). Seawater was collected outside Arcachon Bay (SW France) near the Banc d’Arguin. Immediately after sampling, seawater was filtered using membrane filter of 0.45 μm and then 0.2 μm (Xilab) to eliminate debris and microorganisms. FSW was stored at 4 °C in the dark and was used within 3 days. A few hours before the experiment, FSW was filtered again at 0.2 μm. This reference water was chemically analyzed for to determine pollutant concentration

The range of test concentrations was chosen on the basis of preliminary studies (Gamain et al. 2016). The metal and pesticide solutions were made up from analytical grade copper sulfate (CuSO4, 5H2O) and S-metolachlor. Working solutions were obtained diluting the stock solutions (100 mg L−1 for copper and 250 mg L−1 for S-metolachlor) in FSW and were chemically analyzed. Two concentrations of exposure were selected for copper: 1 and 10 μg L−1 and three for -S-metolachlor: 10, 100, and 1000 ng L−1. The working solutions were chemically analyzed to confirm pollutant concentrations.

Copper analysis

For chemical analysis of copper at 1 and 10 μg L−1, each seawater sample was acidified with 5% of nitric acid (nitric acid 65%, Fluka). Samples were then analyzed by inductively coupled plasma optic emission spectrometry (ICP-OES, Vista Pro, Agilent Technologies) and by inductively coupled plasma mass spectrometry (ICP-MS, Xseries2, Thermo Fisher Scientific). The standards solutions were prepared from a multi-elementary calibration solution (Astasol-Mix M010, Analytika, Czech Republic), in a seawater certified solution (NASS-6 from NRCC-CNRC, Ottawa, Canada). The samples were diluted in a 3% final nitric acid solution (made from a nitric acid 65% Fisher Scientific Trace Metal Grade solution) 1:2 (v/v) for ICP-OES analysis and 1:3 (v/v) for ICP-MS analysis. Quantification limit were 10 μg L−1 (ICP-OES) and 0.3 μg L−1 (ICP-MS).

Metolachlor analysis

Chemical extraction

Fifty-milliliter water samples (pH adjusted to 7.0 ± 0.1 with HCl 0.1 N) were filtered using GF/F glass microfiber filters (0.7-μm pore size). Before analysis, pre-concentration of the analytes was performed using solid-phase extraction (SPE) with Oasis HLB cartridges (Waters), according to the method described by Lissalde et al. (2011). SPE was conducted using a Visiprep 12-port manifold (Supelco, France). The conditioning, extraction, and rising steps were carried out under a 53.33 kPa vacuum. The SPE cartridges were successively washed with 5 mL of methanol, conditioned with 5 mL of ultrapure water, loaded with 50-mL water samples, then rinsed with 5 mL of ultrapure water (UPW) containing 15% HPLC grade methanol. Cartridges were then dried under a nitrogen stream for 30 min. Elutions were achieved with 3 mL of methanol, followed by 3 mL of a mix of methanol/ethyl acetate (75:25 v/v). Then, 2.5 μL of a solution of internal standard (metolachlor d6) at 1 ng μL−1 was then added to the 6-mL extracts, followed by solvent evaporation under a gentle stream of nitrogen and then dissolved in 250 μL of UPW containing 10% HPLC grade acetonitrile prior to analysis.

Instrumentation and data treatment

Metolachlor analyses were performed through liquid chromatography ACQUITY UPLC H-Class coupled to a Xevo G2-S TOF-MS (Waters). The electrospray source was operated in positive mode at 0.7 kV and the sample cone voltage set at 30 V. Nitrogen was used as nebulizer (flow rate 50 L h−1, 150 °C) and desolvation gas (flow rate 1200 L h−1, temperature 600 °C). Data was acquired in the range from 50 to 1200 m/z and acquisition speed was set to 0.2 s. The resolving power full width at half maximum (FWHM) was 30,000 at m/z 556.2771 (leucine encephalin used as lockmass compound).

Data was acquired using MSE in order to obtain both protonated molecular ions at low collision energy (CE = 6 eV) and/or adducts and fragment ions with a collision energy ramp (CE = 10–30 eV). Separation was performed on an ACQUITY BEH C18 column (100 × 2.1 mm, 1.7 μm) from Waters (Milford, MA, USA) with a column temperature of 45 °C and using a binary gradient of water (A) and methanol (B) both containing ammonium acetate (10 mM) at pH 5.0. A flow rate of 0.45 mL min−1 was used and the gradient ranged as follows: 98% A (0–0.25 min), 1% A (12.25–13 min), and 98% A (13.01–17.00). Injection volume was 20 μL. Data treatment was performed with MassLynx v4.1.

Method validation and quality controls

Analytical method was validated in terms of calibration linearity, extraction recoveries, and limits of quantifications (LOQ) according to the French standard NF T90-210. Recovery, LOQ, and LOD are shown in Table 2 for the metolachlor into the different samples. For the quality controls, SPE was routinely controlled, and the recoveries of two levels of spiked mineral water (e.g., 40 and 200 ng L−1) were evaluated for each batch. The periodic control of two calibrating standards (e.g., 2 and 25 μg L−1, every 10 samples) and analytical blanks were performed as well.

Table 2 Acquisition and validation data for the metolachlor

Statistical analysis

All data is expressed as means ± standard deviation (SD). Data was first processed using the transformation: p′ = arc sin \( \sqrt{p} \); p corresponds to raw data (frequency of abnormalities) specified in p values from 0 to 1 (Legendre and Legendre 1998). Homogeneity of variance (Levene’s test) was verified and statistical analysis was performed by the Kruskal-Wallis tests. Differences between data from different conditions were tested using Kruskal post hoc test (equivalent to the Tuckey HSD test for non-parametric data).

Principal component analysis (PCA) was performed to analyze both the spatial distribution pattern of the different studied parameters and the relative relationship between treatment groups. Statistical analyses were performed using Statistica v12 (Statsoft, Maisons-Alfort, France).


Chemical analysis

The copper and S-metolachlor concentrations measured in the reference seawater and the different working solutions are shown in Table 3. These analyses revealed the presence of copper in the reference seawater at concentrations of 2.63 μg L−1. At the nominal concentration of 1 μg L−1, measured concentration was increased of 1.1 μg L−1 compared to control (3.75 μg L−1). At 10 μg L−1, the measured concentration was 18% above expected concentration. Analyses revealed the presence of S-metolachlor in the reference seawater at concentrations lower than 5 ng L−1. At the nominal concentration of 10 ng L−1, measured concentration was 18 ng L−1. At 100 and 1000 ng L−1, measured concentrations were closed to expected, one with a range of variation of the order of 8.1% and 8.4% respectively.

Table 3 Nominal and measured copper and S-metolachlor concentrations (mean values ± SD) at the beginning of the embryotoxicity test

Swimming activity

In control, the average speed of larvae was 144.4 ± 34.6 μm s−1 and the maximum speed was 297.3 ± 85.2 μm s−1 (Fig. 4) while the trajectory was mainly rectilinear (80.8 ± 15.6% of larvae) and to a lesser extent, circular (17.5 ± 13.3%) (Fig. 5). Motionless larvae were observed little in the control group (1.9 ± 3.1%). In the presence of copper, no significant difference in maximum or average larval speed was observed at concentrations of 1 μg L−1 or 10 μg L−1 compared to the control condition (Fig. 4a). After exposure to S-metolachlor, no significant difference in maximum or average larval speed was observed at both concentrations tested compared to the control condition (Fig. 4b). Following copper exposure at both concentrations, the percentage of circular trajectory increased significantly, while rectilinear swimming larvae significantly declined and a dose-dependent effect was observed (Fig. 5a). In the presence of S-metolachlor, the percentage of circular trajectory tended to increase with increasing S-metolachlor concentrations but it was only significant at 100 ng L−1 (Fig. 5b). In the meanwhile, rectilinear trajectory percentage was reduced at all concentrations tested but it was only significant at 10 and 1000 ng L−1.

Fig. 4

Average and maximum speed (μm s−1) of D-larvae exposed to different concentrations of a copper (μg L−1) or b S-metolachlor (ng L−1). Data points represent mean ± SD with 3–5 replicates per concentration and 12–74 larvae for each replicate

Fig. 5

Trajectories of D-larvae exposed to different concentrations of a copper (μg L−1) or b S-metolachlor (ng L−1). Data points represent mean ± SD with 3–5 replicates per concentration and 12–74 larvae for each replicate. Different letters indicated significant differences between exposure conditions (p < 0.05)


Abnormal larvae average frequency was 17.0 ± 2.4% in control condition (Fig. 6 and Table 4). The percentage of developmental arrest increased significantly at 10 μg L−1 of copper while a dose-dependent increase of malformed larvae was observed at all copper concentrations tested (Fig. 6a and Table 4). After exposure to S-metolachlor, the percentage of developmental arrest was significantly increased from 100 ng L−1 whereas the percentage of abnormalities increased significantly only at 1000 ng L−1 (Fig. 6b and Table 4).

Fig. 6

Percentages of developmental arrest (DA) and abnormal D-larvae exposed to different concentrations of a cooper (μg L−1) or b S-metolachlor (ng L−1). Data points represent mean ± SD. N = 4 replicates per condition, and 100 larvae for each replicate. Different letters indicated significant differences between exposure conditions (p < 0.05)

Table 4 Abnormal larvae and malformation frequency (%) in oyster larvae exposed to copper or S-metolachlor (Mean values ± SD)

Relationship between malformations and swimming behavior

PCA featured 60.88% of the selected variables considering the two first axes. The first axis accounted for 38.23%, while the second axis accounted for 22.65% of the variability.

For the plot of variables (Fig. 7a), the PC1 was positively loaded by both circular trajectories and shell abnormalities and negatively loaded by rectilinear trajectories. PC2 was negatively loaded by maximum and average speeds. Rectilinear trajectories appeared positively correlated to normal larvae (r = 0.47, p = 0.012) while larvae with shell anomalies were positively correlated to circular trajectories (r = 0.60, p = 0.001). In addition, motionless larvae were obviously anti-correlated to maximum and average speeds.

Fig. 7

Principal component analysis of the different types of abnormalities and trajectories of D-larvae. Analysis represents normalized coefficients on the first two axes (axis 1 = 38.23%; axis 2 = 22.65%) for three D-larvae abnormalities (normal, shell abnormalities, mantle abnormalities), three trajectories (R = rectilinear; S = motionless, C = circular) two speed (Vit Moy: average speed, Vit Max: maximum speed). The variables factor map is shown (a) and the individual factor map in (b). Pearson correlation analysis between shell abnormalities and circular trajectories (c) and rectilinear trajectories and normal D-larvae (d)

For the plot of treatment groups (Fig. 7b), the PC1 and PC2 allowed a good separation of the control group versus copper 1 μg/L and copper 10 μg/L groups. The separation of the control group and metolachlor 100 ng/L and metolachlor 1000 ng/L groups was well defined. In contrast, the metolachlor 10 ng/L group was partly overlapping with the control and metolachlor 100 and 1000 ng/L groups. Mantle malformations appear positively correlated with exposure to copper at 10 μg/L and S-metalchlor at 1000 ng/L.


Normal swimming behavior of oyster D-larvae

The video tracking system described in this paper represents an operational tool to evaluate different swimming parameters exhibited by D-larvae of the Pacific oyster C. gigas. A lot of other automated systems have been described to analyze swimming activity, motility or frequency of pulsation in numerous invertebrates (for review, see Faimali et al. 2017) but to our knowledge, this study is the only one analyzing both swimming speed and trajectory.

Our semi-automated video tracking system was able to identify three major types of larval path: rectilinear, circular, and motionless. It is to be noted that these do only consider two dimensions and that larvae oysters moving in their environment can move in three dimensions. A study in three dimensions would not affect rectilinear paths, but circular paths might split between real circles and 3D spirals.

In our experiment, under optimal development conditions (reference filtrated seawater, 24 °C, salinity of 33 usi), oyster larvae mainly adopted rectilinear trajectories (81% ± 15). Thus, the most common behavior of the larvae would correspond to rectilinear swimming. These results confirm the early work of His et al. (1999) on the mussel Mytilus galloprovincialis, which suggested that a circular or spiral swimming denoted erratic larval behavior.

In the same laboratory conditions (24 °C, salinity of 33 usi), the average speed and mean maximum velocity of oyster D-larvae were 144 and 297 μm s−1, respectively. Suquet et al. (2012) reported an average swimming speed of 105 μm s−1 for C. gigas D-larvae in the absence of contamination. This value is slightly lower than that calculated in our study. This difference can be explained, at least in part, by the temperature difference between the two experiments, e.g., 19 °C for Suquet’s study and 24 °C for this study. The 24 °C correspond to the ideal temperature of development for C. gigas (AFNOR 2009; Gamain et al. 2016). Incubation of C. gigas embryos at temperature less than or equal to 20 °C give rise to significant increase of developmental abnormalities (Gamain et al. 2016). Those abnormalities have likely consequences on the behavior and the speed of swimming of larvae. Indeed, Horiguchi et al. (1998) reported unusual swimming behavior, low swimming activity, and irregular movement of cilia due to atrophy of velum in larvae of different mollusk species exposed to organotin compounds.

Effects of model pollutants on swimming behavior of oyster D-larvae

In our experimental setup (24 °C, salinity of 33 usi), there are no significant differences in average and maximum swimming speeds up to 10 μg/L of copper and 1 μg/L of S-metolachlor. In contrast, the percentage of rectilinear larval paths decreased with the concentration of copper and S-metolachlor (except for the 1000 ng/L) while the percentage of circular larval paths increases with the concentration of copper and S-metolachlor (except for the 1000 ng/L). The absence of effects on swimming speed means that the two compounds did not target the metabolism and/or the neuro-muscular system of oyster D-larvae. Besides, a significant change in larval trajectory was observed at the lowest tested concentrations of copper (1 μg/L) and S-metolachlor (10 ng/L) with a striking shift of rectilinear paths to circular ones. This trajectory shift was likely associated to the appearance of morphological anomalies in developing oyster embryos since a significant positive correlation was noted between circular trajectory and shell abnormality of D-larva.

No data is available about the effects of S-metolachlor on swimming activity among aquatic organisms. Besides, a few works have documented effects of copper on fish and invertebrates. LaBreche et al. (2002) observed a decrease of motility in larvae of Mercenaria mercenaria exposed to concentration of Cu above 10 μg/L. In nauplii of Artemia salina a decrease of mobility was detected from 1 mg/L of copper (Kokkali et al. 2011). Kwok and Ang (2013) also reported an inhibition of swimming activity of larvae of the coral Platygyra acuta at a concentration above 40 μg/L. After 48 h of depuration, the larval motility was not recovered, indicating a possible persistent effect of copper.

The toxicity of copper and S-metolachlor for early life stage of aquatic species is well-documented. It has been shown that both compounds induced significant developmental anomalies and DNA damage in bivalves at environmentally realistic concentrations (Mai et al. 2012, 2013; Gamain et al. 2017a). Behavior is an integrated and whole-organism response (Faimali et al. 2017). Occasional and/or limited changes in this behavior are assumed to be an immediate and adaptive response to stress. On the other hand, permanent and deep disruption to behavior is the result of irreversible changes at the sub-individual level that can have a dramatical impact on fitness and survival of individuals. The significant positive correlation between circular trajectory and anomalies of the shell in oyster larvae exposed to copper or S-metolachlor demonstrate the deep impact of these molecules on the embryo-larval development of this specie. This study demonstrates, once again, the very high sensitivity of the early development stages of oyster to chemical contamination.

Application of swimming activity monitoring of oyster larvae

For over a decade, researchers have been using swimming activity in aquatic invertebrates as new endpoints to evaluate toxicity of chemicals (Garaventa et al. 2010; Chevalier et al. 2015; Faimali et al. 2017). This kind of endpoint is increasingly used for environmental risk or water quality assessment in non-model organisms such as crustaceans and sea urchins (Garaventa et al. 2010; Morgana et al. 2016). Although mussels and oysters are widely used as sentinel species for pollution monitoring, to date, very few works have focused on the swimming behavior of bivalves exposed to pollutants (Horiguchi et al. 1998).

In the present study, the effects of pollutants were clearly detected on swimming behavior of oyster larvae at concentrations as low as 1 μg/L of copper and 10 ng/L of S-metochlor. These levels of contamination are environmentally relevant and are currently detected in the seawater of European coastal areas, notably in the Arcachon Bay (Gamain et al. 2017b). The swimming activity of oyster larvae could be used as an early warning indicator of the quality of water and/or of the toxicity of pollutants.

Abnormal swimming behavior in D-larvae can directly affect their survival or fitness. In addition, abnormal trajectories can impair dispersion of larvae and colonization of new habitats. Both phenomena could impair recruitment and at terms impact survival of wild population and viability of production of oysters. In this respect, swimming behavior could be used as a suitable tool to analyze the health status of larvae from wild or farmed oyster populations.

In the future, we can expect the use of video-tracking of oyster larvae to become a convenient tool to help shellfish farmers choose optimum batches of larvae and to aid environmental agencies in assessing coastal water quality.


This study describes a semi-automated image analysis tool to analyze bivalve D-larvae swimming behavior. Oyster larvae exposed to environmental concentrations of copper or S-metolachlor exhibited disrupted swimming trajectories, but average and maximum speeds were not impacted. Rectilinear trajectories appeared positively correlated to normal larvae, while larvae with shell anomalies were positively correlated to circular trajectories. This study showed that video tracking of oyster swimming behavior is a sensitive and easy to perform tool for ecotoxicological testing. Using this tool at a larger scale could call for some improvements, particularly in terms of temperature control and automatic trajectory reading.


  1. AFNOR XP T90-382 (2009) Qualité de l'eau - Bio indicateur de la toxicité potentielle de milieux aqueux - Détermination de la toxicité potentielle d'échantillons aqueux sur le développement embryo-larvaire de bivalve. p. 19

  2. Auby I, Bocquene G, Quiniou F, Dreno JP (2007) Etat de la contamination du Bassin d’Arcachon par les insecticides et les herbicides sur la période 2005-2006. Impact Environn 24:108

    Google Scholar 

  3. Auby I, Maurer D, Passoni S, Heroin D, Rigouin L, Méteigner C, Perrière-Rumèbe M, Tournaire M-P (2014) Reproduction de l’huître creuse dans le Bassin d'Arcachon Année 2014. Ifremer 55.

  4. Caldwell CA, Jacobi GZ, Anderson MC, Parmenter RR, McGann J, Gould WR, DuBey R, Jacobi MD (2013) Prescribed-fire effects on an aquatic community of a Southwest Montane Grassland System. N Am J Fish Manag 33(5):1049–1062

    Article  Google Scholar 

  5. Chevalier J, Harscoët E, Keller M, Pandard P, Cachot J, Grote M (2015) Exploration of Daphnia behavioral effect profiles induced by a broad range of toxicants with different modes of action. Environ Toxicol Chem 34:1760–1769

    CAS  Article  Google Scholar 

  6. Faimali M, Gambardella C, Costa E, Piazza V, Morgana S, Estevez-Calvar N, Garaventa F (2017) Old model organisms and new behavioral end-points: swimming alteration as an ecotoxicological response. Mar Environ Res 128:36–45

    CAS  Article  Google Scholar 

  7. FAO (2019) Fisheries and aquaculture Department. Fish Stat Crassostrea gigas (Thunberg, 1793).

  8. Festa RA, Thiele DJ (2011) Copper: an essential metal in biology. Curr Biol 21:R877–R883

    CAS  Article  Google Scholar 

  9. Flemming CA, Trevors JT (1989) Copper toxicity and chemistry in the environment: a review. Water Air Soil Pollut 44:143–158

    CAS  Article  Google Scholar 

  10. Gamain P, Gonzalez P, Cachot J, Pardon P, Tapie N, Gourves PY, Budzinski H, Morin B (2016) Combined effects of pollutants and salinity on embryo-larval development of the Pacific oyster, Crassostrea gigas. Mar Environ Res 113:31–38

    CAS  Article  Google Scholar 

  11. Gamain P, Gonzalez P, Cachot J, Clérandeau C, Mazzella N, Gourves P-Y, Morin B (2017a) Combined effects of temperature and copper and S-metolachlor on embryo-larval development of the Pacific oyster, Crassostrea gigas. Mar Pollut Bull 115:201–210

    CAS  Article  Google Scholar 

  12. Gamain P, Cachot J, Gonzalez P, Budzinski H, Gourves P-Y, Morin B (2017b) Do temporal and spatial parameters or lifestyle of the Pacific oyster Crasssostrea gigas affect pollutant bioaccumulation, offspring development and tolerance to pollutants? Front Mar Sci 4:58.

    Article  Google Scholar 

  13. Garaventa F, Gambardella C, Di Fino A, Pittore M, Faimali M (2010) Swimming speed alteration of Artemia sp. and Brachionus plicatilis as a sub-lethal behavioural end-point for ecotoxicological survey. Ecotoxicology 19:512–519

    CAS  Article  Google Scholar 

  14. Handy RD, Sims DW, Giles A, Campbell HA, Musonda MM (1999) Metabolic trade-off between locomotion and detoxification for maintenance of blood chemistry and growth parameters by rainbow trout (Oncorhynchus mykiss) during chronic dietary exposure to copper. Aquat Toxicol 47:23–41

    CAS  Article  Google Scholar 

  15. His E, Beiras R, Seaman MNL (1999) The assessment of marine pollution bioassays with bivalve embryos and larvae. Adv Mar Biol 37:1–178

    Article  Google Scholar 

  16. Horiguchi T, Imai T, Cho HS, Shiraishi H, Shibata Y, Morita M, Shimizu M (1998) Acute toxicity of organotin compounds to the larvae of the rock shell, Thais clavigera, the disk abalone, Haliotis discus discus and the giant abalone, Haliotis madaka. Mar Environ Res 46:469–473

    CAS  Article  Google Scholar 

  17. Kapur JN, Sahoo PK, Wong AKC (1985) A new method for gray-level picture thresholding using the entropy of the histogram. Computer Vis Graph Image Process 29:273–285

  18. Kazlauskiene N, Svecevičius G, Petrauskiene L, Vosyliene MZ (2010) Behavioural responses of medicinal leech and rainbow trout exposed to crude oil and heavy fuel oil in ontogenesis. Pol J Environ Stud 19:429–433

    CAS  Google Scholar 

  19. Kokkali V, Katramados I, Newman JD (2011) Monitoring the effect of metal ions on the mobility of Artemia salina Nauplii. Biosensors 1:36–45

    CAS  Article  Google Scholar 

  20. Kwok CK, Ang PO (2013) Inhibition of larval swimming activity of the coral (Platygyra acuta) by interactive thermal and chemical stresses. Mar Pollut Bull 74:264–273

  21. LaBreche TMC, Dietrich AM, Gallagher DL, Shepherd N (2002) Copper toxicity to larval Mercenaria mercenaria (hard clam). Environ Toxicol Chem 21:760–766

    CAS  Article  Google Scholar 

  22. Le Bihanic F, Sommard V, de Lansalut P, Anaïk P, Grasset J, Berrada S, Budzinski H, Cousin X, Morin B, Cachot J (2015) Environmental concentrations of benz[a]anthracene induce developmental defects and DNA damage and impair photomotor response in Japanese medaka larvae. Ecotox Environ Safe 113:321–328

  23. Legendre P, Legendre L (1998) Numerical ecology. Transformation:1–4

  24. Lissalde S, Mazzella N, Fauvelle V, Delmas F, Mazellier P, Legube B (2011) Liquid chromatography coupled with tandem mass spectrometry method for thirty-three pesticides in natural water and comparison of performance between classical solid phase extraction and passive sampling approaches. J Chromatogr A 1218(11):1492–1502

    CAS  Article  Google Scholar 

  25. Lv Z, Tek A, Da Silva F, Empereur-mot C, Chavent M, Baaden M (2013) Game on, science - how video game technology may help biologists tackle visualization challenges. Plos One.

  26. Mai H, Cachot J, Brune J, Geffard O, Belles A, Budzinski H, Morin B (2012) Embryotoxic and genotoxic effects of heavy metals and pesticides on early life stages of Pacific oyster (Crassostrea gigas). Mar Pollut Bull 64:2663–2670

    CAS  Article  Google Scholar 

  27. Mai H, Morin B, Pardon P, Gonzalez P, Budzinski H, Cachot J (2013) Environmental concentrations of irgarol, diuron and S-metolachlor induce deleterious effects on gametes and embryos of the Pacific oyster, Crassostrea gigas. Mar Environ Res 89:1–8

    CAS  Article  Google Scholar 

  28. Martin JM, Saaristo M, Bertram MG, Lewis PJ, Coggan TL, Clarke BO, Wong BBM (2017) The psychoactive pollutant fluoxetine compromises antipredator behaviour in fish. Environ Pollut 222:592–599

    CAS  Article  Google Scholar 

  29. McCallum ES, Krutzelmann E, Brodin T, Fick J, Sundelin A, Balshine S (2017) Exposure to wastewater effluent affects fish behaviour and tissue-specific uptake of pharmaceuticals. Sci Total Environ 605-606:578–588

    CAS  Article  Google Scholar 

  30. Melvin SD, Wilson SP (2013) The utility of behavioral studies for aquatic toxicology testing: a meta-analysis. Chemosphere 93:2217–2223

    CAS  Article  Google Scholar 

  31. Mileikovsky SA (1973) Speed of active movement of pelagic larvae of marine bottom invertebrates and their ability to regulate their vertical position. Mar Biol 23:11–17

    Article  Google Scholar 

  32. Morgana S, Gambardella C, Falugi C, Pronzato R, Garaventa F, Faimali M (2016) Swimming speed alteration in the early developmental stages of Paracentrotus lividus sea urchin as ecotoxicological endpoint. Mar Environ Res 115:11–19

    CAS  Article  Google Scholar 

  33. Nussbaum-Krammer CI, Neto MF, Brielmann RM, Pedersen JS, Morimoto RI (2015) Investigating the spreading and toxicity of prion-like proteins using the metazoan model organism C. elegans. J Vis Exp (95):e52321

  34. Scott GR, Sloman KA (2004) The effects of environmental pollutants on complex fish behaviour: integrating behavioural and physiological indicators of toxicity. Aquat Toxicol 68:369–392

    CAS  Article  Google Scholar 

  35. Sommers F, Mudrock E, Labenia J, Baldwin D (2016) Effects of salinity on olfactory toxicity and behavioral responses of juvenile salmonids from copper. Aquat Toxicol 175:260–268

    CAS  Article  Google Scholar 

  36. Suquet M, Le Mercier A, Rimon F, Mingant C, Haffray P, Labbe C (2012) Setting tools for the early assessment of the quality of thawed Pacific oyster (Crassostrea gigas) D-larvae. Theriogenology 78:462–467

    CAS  Article  Google Scholar 

  37. Suquet M, Rimond F, Cosson J, Wilson-Leedy J, Lebrun L, Queau I, Mingant C, Fauve C (2013) Effect of age and environmental conditions on the movement characteristics of Pacific oyster (Crassostrea gigas) trochophores. J Appl Ichthyol 29:1145–1148

    Article  Google Scholar 

  38. Tapie N, Chevance DL, Budzinski H (2017) Repar pesticides : quantification de la présence dans les eaux, année 2012, 22 p.

  39. Troost K, Veldhuizen R, Stamhuis EJ, Wolff WJ (2008) Can bivalve veligers escape feeding currents of adult bivalves? J Exp Mar Biol Ecol 358:185–196

    Article  Google Scholar 

  40. Turner A (2010) Marine pollution from antifouling paint particles. Mar Pollut Bull 60:159–171

    CAS  Article  Google Scholar 

  41. Van der Schalie WH, Shedd TR, Knechtges PL, Widder MW (2001) Using higher organisms in biological early warning systems for real-time toxicity detection. Biosens Bioelectron 16:457–465

    Article  Google Scholar 

Download references


The authors thank the Aquitaine Region (OSQUAR Project), CPER A2E, Intermunicipal Union of Arcachon Bay (SIBA) and Water Agency Adour Garonne (AEAG) for their financial support. This work was part of the LABEX COTE cluster of excellence “Continental To coastal Ecosystems: evolution, adaptability and governance”. We thank James Emery for providing English proofreading.

Author information



Corresponding author

Correspondence to Jérôme Cachot.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Responsible editor: Philippe Garrigues

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Gamain, P., Roméro-Ramirez, A., Gonzalez, P. et al. Assessment of swimming behavior of the Pacific oyster D-larvae (Crassostrea gigas) following exposure to model pollutants. Environ Sci Pollut Res 27, 3675–3685 (2020).

Download citation


  • Copper
  • Early life stage
  • Image analysis
  • Malformations
  • S-metolachlor
  • Speed
  • Trajectory