Eastern oyster ( Crassostrea virginica ) nursery production in tropical coastal lagoons in Yucatán, Mexico: nonlinear regression modeling and relationships with environmental variables

. Nonlinear regression modeling was used to study the nursery production of the oyster Crassostrea virginica in the Celestún (CL) and Rio Lagartos (RL) lagoons in Yucatán, Mexico. Relationships between production parameters and environmental variables were also established. Spat (2.40 ± 0.20 mm) was obtained from a hatchery and reared in Nestier-type trays using an off-bottom system. Oyster cultivation took place from May 2021 to September 2021 and ended after 122–126 days when oysters reached 30.00 mm. There were no significant differences in final height (CL: 30.80 ± 0.42 mm; RL: 31.80 ± 0.65 mm; P = 0.18) and growth rate (CL: 0.23 ± 0.02 mm·d –1 ; RL: 0.23 ± 0.01 mm·d –1 ; P = 0.98). Final survival was 71.45% in CL and 99.40% in RL. Nonlinear regression curves were statistically satisfactory for analyzing growth and survival. Except for dissolved oxygen, mean salinity (CL: 15.23 ppt; RL: 35.02 ppt), temperature (CL: 29.64 °C; RL: 31.02 °C), dissolved oxygen (CL: 4.50 mg·L –1 ; RL: 5.04 mg·L –1 ), pH (CL: 8.10; RL: 8.34), chlorophyll a (Chl a ) (CL: 3.23 mg·m –3 ; RL: 6.85 mg·m –3 ), and total dissolved solids (CL: 16,101 mg·L –1 ; RL: 34,838 mg·L –1 ) were significantly higher in RL ( P < 0.05). Except for Chl a , the environmental variables were more stable in RL than in CL ( P < 0.05). In RL, the growth rate was positively related to salinity and pH. In CL, the growth rate slowed when salinity decreased, and the mortality rate diminished when salinity, dissolved oxygen, and total dissolved solids increased and pH decreased. Salinity was mainly responsible for the observed differences in production between lagoons. Even when temperature and salinity were high in RL, acceptable growth rate and survival were observed, possibly due to stable rearing conditions.


Introduction
The Eastern Oyster (Crassostrea virginica) has proven to be suitable for farming, mainly in the coastal waters of Canada and the United States of America (USA) (Poirier et al. 2020, Bodenstein et al. 2023).Several studies have been conducted on C. virginica farming, addressing issues such as alternative rearing gears (Mallet et al. 2013, Thomas et al. 2019, Grizzle et al. 2020) and the effects of desiccation and tumbling practices (Bodenstein et al. 2021).
In Mexico, C. virginica farming began in 2010 after a hatchery (Centro Ostrícola Tecnológico de Tabasco) commenced operations in the state of Tabasco with a production capacity of 40 million (M) spats per month (pers. comm., Tordecillas-Guillén 2023).Currently, 27 farms operate in Tabasco, and 12 farms operate in the state of Tamaulipas (CONAPESCA 2022).Despite the proximity of the hatchery to other states, such as Yucatán, there are no reports of commercial oyster aquaculture in other states.In addition, few antecedents are available regarding C. virginica farming in Yucatán.Rihani et al. (1989) reported the results of the first C. virginica experimental cultivation assays using juveniles from Tabasco.Cabrera-Rodríguez et al. (1997) investigated the feasibility of oyster rearing in the Rio Lagartos Lagoon.Vera and Aldana (2000) evaluated cultivating oysters in the water supply channels of a shrimp farm.
Recent studies have focused on the influence of environmental variables on C. virginica growth and survival.La Peyre et al. (2013) assessed recruitment, growth, and mortality along a salinity gradient in an estuarine zone in Louisiana, USA.Rybovich et al. (2016) examined the effects of low salinity and high temperature on growth and survival.McFarland et al. (2022) evaluated osmotic limits at moderate and elevated temperatures during different life stages.Few studies have focused on C. virginica during the nursery stage.Among these, Bishop and Hooper (2005) compared C. virginica growth and mortality rates during the nursery stage.Poirier et al. (2020) tested the efficacy of a bouncing-bucket system to improve oyster shell strength and thickness.
The present study evaluated the biotechnical viability of C. virginica production when reared off-bottom during the nursery stage in 2 tropical lagoons in Yucatán, Mexico.In addition, the relationships between growth and survival parameters and the environmental conditions of the lagoons were analyzed.To our knowledge, this is the first study to address this topic.
Celestún Lagoon is a karstic shallow estuary that contains limestone; freshwater entry via groundwater discharge into the lagoon varies according to the rainfall regime (Morelos- Villegas et al. 2018).Rio Lagartos is the largest coastal lagoon in Yucatán and exhibits hypersaline conditions (>40 ppt) during most of the year (Vega-Cendejas et al. 2004).The region has a well-defined weather regime with 3 seasons: dry (March to May), rainy (June to October), and windy (November to February); strong winds and temperatures below 22 °C are present during the windy season (Morelos- Villegas et al. 2018).

Nursery description
The nursery phase lasted 122 days in CL (20 May 2021 to 19 September 2021) and 126 days in RL (21 May 2021 to 24 September 2021), during which the oysters attained a mean shell height (distance between the hinge and the opposite margin) of at least 30 mm.Spat of C. virginica with a mean shell height of 2.4 ± 0.2 mm were obtained from the commercial hatchery Centro Ostrícola Tecnológico de Tabasco.Rearing was carried out in Nestier-type trays (DM Plast, Jalisco, Mexico).The oysters were stocked in plastic mesh bags (55 × 35 cm; mesh size: 0.8 mm); 5,000 individuals were stocked per bag.The bags were placed in the Nestier trays to form modules (3 vertically stacked trays per module) which were attached off-bottom to 30-m longlines.The initial oyster populations were 100,000 (CL) and 50,000 (RL) individuals.
The oysters were transferred to larger plastic bags (55 × 55 cm; mesh size: 1.35 mm) after reaching a shell height of >10 mm.When the oysters attained a shell height of >20 mm, they were removed from the bags and placed directly into the trays, using 50% of the bottom area until the end of the trials.This protocol increased the modules to a maximum of 42 at CL and 18 at RL. Fouling organisms were removed from the bags and trays each week by scrubbing with a stiff hand brush.

Data collection and processing
Samples of 30 individual oysters were randomly selected every 7-14 days to measure shell height using a dial caliper (505-730 D15TX; Mitutoyo, Kanagawa, Japan) (precision: 0.02 mm).The dead oysters (identified as empty shells or gaping oysters) were counted and removed from the bags or trays each week.Temperature (°C), salinity (ppt), dissolved oxygen (mg•L -1 ), pH, and total dissolved solids (mg•L -1 ) were recorded using a handheld multimeter (Pro-Plus; Yellow Spring Instruments, Yellow Springs, USA).Chlorophyll a data were obtained from the Giovanni platform of the National Aeronautics and Space Administration (NASA, 2022).
Daily growth and mortality rates were calculated for the periods between successive sampling dates as the differences in height and survival (%) divided by the number of days between sampling dates.The values of environmental variables corresponding to the successive sampling dates were averaged to represent the prevailing conditions.

Growth and survival analyses
Equations (1) and (2) were used to analyze oyster growth as a function of time (t).Equation ( 1) is the curve proposed by Estrada-Pérez et al. (2018): where h t is mean height, H i is the initial mean height, b is the growth rate, C is the amplitude of height variation (deviation) around a linear growth trend, L is the length of the period between minima or maxima deviations, and S is a time adjustment parameter.Equation (2) was proposed by Serna-Gallo et al. (2014): where h t is mean height, H 0 is the initial mean height, H f is the mean height at the end of the nursery stage, k is a parameter representing the rate at which the height changes from its initial value to its final value, t is elapsed time units, and c is the time units elapsed at the end of the rearing period.
Oyster survival in CL was analyzed as a function of time (t), using the following model: where S t is the percentage of survivors, S i is the initial survival of the stock (i.e., 100%), A is the initial instantaneous mortality rate, and B is the rate at which the instantaneous mortality rate increases.Oyster mortality in RL was negligible; thus, statistical analysis was not required.

Environmental variables and relationships with growth and mortality rates
The mean values of the environmental variables were compared to test differences in environmental rearing conditions between the lagoons.The relative stability of the environmental conditions within and between the lagoons was analyzed by comparing the coefficient of variation (i.e., ratio of standard deviation to the variable mean).Subsequently, linear regression analyses were conducted to evaluate possible relationships between environmental variables and growth and mortality rates.
A relationship was established between the coefficients of variation of oyster height (CV h ) and salinity (CV s ) in CL: where CV 0 is the coefficient of variation of the growth rate when the coefficient of variation of salinity is zero, and G is the instantaneous rate at which CV h changes as a function of CV s .

Statistical analyses
Equations ( 1), ( 2), (3), and (4) were estimated using nonlinear regression fitting methods.Equations ( 1) and (2) were fitted to the height data collected in both lagoons; the best fit was determined using the Akaike information criterion (AIC).D' Agostino-Pearson and Shapiro-Wilk tests were used to assess the normality of the regression residuals.The lack of fit of the model and parameter skewness were also tested.Parameter skewness was measured using the Hougaard index (h-index) and compared with the index values recommended by Ratkowsky (1983).Accordingly, h-index absolute values were categorized as follows: ideal (<0.10) (almost linear), adequate (0.10-0.25), noticeably skewed (0.25-1.00), and considerably skewed (>1).
The final mean oyster heights of the lagoons were compared with a one-way analysis of variance (ANOVA), considering that the difference in the nursery length between the sites was negligible (i.e., 3.2%).A repeated measures ANOVA was used to compare the growth rate values and environmental variables of the lagoons.Data normality was evaluated with the Shapiro-Wilk test before conducting the analyses.A non-parametric Mann-Whitney test was used when the data were not normally distributed.The coefficients of variation of  the environmental variables were compared using the method proposed by Forkman (2009).Statistica v. 6.1 (StatSoft, Inc., Tulsa, USA), GraphPad Prism v. 9.0 (GraphPad Software, San Diego, USA), and MedCalc Statistical Software v. 20.218 (MedCalc Software Ltd., Ostend, Belgium) were used to conduct the statistical analyses.Significance was evaluated at α = 0.05.
According to the AIC, the Estrada-Perez et al. ( 2018) model showed a better fit to the oyster growth observed in CL, while the Serna-Gallo et al. ( 2014) equation resulted in a better fit for the RL data (Table 1).It is worth noting that the Estrada-Perez et al. ( 2018) model showed erratic behavior when fitting growth data for RL, yielding unrealistic values of 130 days for parameter L (the length of the period between minimum or maximum height deviations), which is longer than the nursery period.Erratic behavior was also evident by the negative growth rate values (parameter B).
The variation in the growth rate observed in CL (Fig. 2, Table 2) throughout the nursery phase was adequately described by the model of Estrada-Pérez et al. (2018).The growth model proposed by Serna-Gallo (2014) adequately described oyster growth in RL (Fig. 2, Table 2).Final oyster survival was 71.45% in CL.In RL, mortality was negligible, and survival was as high as 99.40%.The survival model (Eq.3) adequately described the mortality observed in CL (Fig. 3, Table 3).
The values of environmental variables recorded in the lagoons during the nursery phase are shown in Figures 4 and  5. Except for mean dissolved oxygen, which was not significantly different between sites, the values of the environmental variables were significantly higher in RL (Supplementary Material Table S1).The coefficients of variation of the environmental variables showed that CL was more unstable in terms of salinity and total dissolved solids, while Chl a was more unstable in RL (Supplementary Material Table S1).When compared within each lagoon, salinity (CL) and Chl a (RL) were significantly more unstable than the rest of the variables, while temperature was the most stable variable in both lagoons (Supplementary Material Table S2).
No significant relationships were detected between the growth rate and environmental variables in CL.Nevertheless, 4 periods were identified where oyster height increased or stabilized, which corresponded to periods when salinity was relatively stable (periods I and III) or decreasing (periods II and IV), respectively (Fig. 6).The coefficients of variation during the periods were used to estimate the relative variability of height and salinity.Low CV values indicated high stability, while high values corresponded to periods when the variables were unstable (i.e., when height increased or salinity decreased).A significant and inverse relationship between the CV values of both variables was determined by fitting Eq. (4) (Fig. 7).The results from the regression analysis were: CV 0 = 2.16 ± 0.61, P = 0.07, and h-index = 1.23;G = 7.46 ± 1.47, P = 0.04, and h-index = -1.61(regression: P = 0.01; normality: P = 0.75; lack of fit: P > 0.99).Thus, oyster growth slowed down as salinity decreased.
The results from the linear regression analyses showed that in CL, the mortality rate was inversely related to salinity, dissolved oxygen, and total dissolved solids and positively related to pH.In contrast, in RL, the growth rate was positively related to salinity and pH (Table 4).The slope values in the equations indicated a change in the growth and mortality rates per unitary change in the corresponding environmental variable.For example, an increase (or decrease) of 1 ppt in salinity corresponded to an increase (or decrease) in the daily mortality rate in CL of 0.051%.

Discussion
The coastal lagoons analyzed in the present study showed substantial differences in environmental conditions that are relevant for the nursery production of C. virginica.Noticeable differences were observed in the magnitude, relative stability, and dynamics of the environmental variables between the lagoons.Unstable salinity caused an oscillating oyster growth pattern in CL.At the same time, higher general stability favored more consistent growth in RL (Fig. 2).Oyster height was greater in CL during most of the rearing period, although by the end of the nursery phase, mean height was similar in both lagoons.Despite the difference in growth patterns, the daily growth rates and final heights did not differ significantly.In contrast, a substantial difference was observed in oyster survival; the prevailing conditions in CL were more unfavorable than in RL.
The main factors determining oyster rearing conditions were the water exchange regimes between the lagoons, sea, and groundwater sources.Celestún Lagoon is a shallow estuary with freshwater contributions from groundwater and  Bishop and Hooper (2005) used upwellers to study growth during nursery rearing, allowing for a growth rate of 6.80 mm•month -1 to be estimated.A growth rate of 6.60 mm•month -1 has also been reported for the nursery stage using "Australian-type" cages (pers.comm., Tordecillas-Guillén 2023).These growth rate estimates are similar to those obtained in this study (6.96 mm•month -1 and 7.00 mm•month -1 ).
The optimal water temperature for C. virginica ranges from 20 °C to 32 °C (Lowe et al. 2017).Overall, the temperature values observed in this study were adequate in both lagoons, with the mean value in RL (31.02 °C) nearing the maximum tolerable limit.Information on tolerable and optimal salinity levels for C. virginica is abundant, particularly regarding its wild populations.Tolerable limits range from 10 ppt to 30 ppt, while values within approximately 15-23 ppt are usually considered optimal (Shumway 1996).In the present study, salinity levels of 15.23 ppt were observed in CL, nearing the minimum acceptable limit.In contrast, salinities higher than those considered adequate were observed in RL (35.02 ppt).
Dissolved oxygen higher than 3.00 mg•L -1 is recommended for C. virginica (Baker andMann 1992, Coxe et al. 2023).Acceptable dissolved oxygen levels were observed in the present study, except for CL, where an unacceptably low concentration (2.95 mg•L -1 ; Fig. 4) was recorded at the end of the trial, although this situation prevailed for a short period and did not lead to hypoxic conditions (i.e., <2 mg•L -1 , Coxe et al. 2023).The pH values recorded in the present study (8.10 and 8.30) were within limits considered adequate for C. virginica embryos, larvae, and spat (i.e., 6.80-8.80)(Calabrese andDavis 1966, Beniash et al. 2010).According to Snyder et al. (2017), the ideal chlorophyll concentration for C. virginica farming is 3-10 mg•m -3 .According to this ranking, the mean concentrations of Chl a in the present study were ideal (3.23 mg•m -3 in CL and 6.85 mg•m -3 in RL).No recommended values of   total dissolved solids for C. virginica were found in the reviewed literature.
The salinity tolerance limits of the C. virginica may vary depending on the magnitude and duration of exposures to extreme conditions (McFarland et al. 2013) or sudden changes in salinity (McFarland et al. 2022).In response to extreme changes in salinity, oysters close their valves and seal themselves within their shells to avoid osmotic stress (Rybovich et al. 2016); consequently, oysters stop feeding, and growth rates decrease (Lavaud et al. 2017).In CL, salinity abruptly changed from 24 ppt to 11 ppt from day 28 to day 56 and from 19 ppt to 5.40 ppt from day 98 to day 122 (Fig. 6).During those periods, the growth rate decreased from 0.35 mm•d -1 to 0.01 mm•d -1 and from 0.33 mm•d -1 to 0.03 mm•d -1 .As has been reported by other authors, the growth decreases in this study were most likely due to limited filtration due to the shell being closed.This situation resulted in a pattern in which the growth rate decreased during intervals related to sudden changes in salinity.Later, when salinity stabilized, the growth rate increased again.The decreases and increases in the growth rate observed in CL can thus be attributed to successive periods of stress and no-stress associated with periods of unstable and stable salinity, respectively.
Several studies have indicated that low salinity negatively affects the survival of oyster species.According to Maeda-Martinez et al. (2023), environmental fluctuations, particularly salinity in the range of 32 ppt to 13.6 ppt and high temperatures (>34 °C), caused mortality in farms of Crassostrea corteziensis.Low salinities impair the survival rate of C. virginica at any life stage (La Peyre et al. 2013, Leonhardt et al. 2017).The synergistic effects of salinity and temperature are probably the most important for C. virginica (Heilmayer et al. 2008).Low salinities (<15 ppt) and high water temperatures (>30 °C) have been shown to increase oyster mortality significantly (McFarland et al. 2022).In agreement with those studies, mortality in this study increased in CL after ~42 days (Fig. 3) when salinity fell to 15 ppt, and the temperature was close to 30 °C (Fig. 4).
The nursery phase in CL was conducted during the rainy season, a dynamic period with massive freshwater inputs due to rainfall and groundwater entering the lagoon.According to Herrera-Silveira and Ramírez-Ramírez (1998), CL is characterized by low salinity (2.10-3.70 ppt) and low dissolved oxygen concentrations (<1 mg•L -1 ) during this period, which may explain the negative relationships between the mortality rate and salinity and dissolved oxygen found in the lagoon.The role of pH in oyster mortality is unclear, and the significant relationship obtained for CL may have resulted from correlations between pH and other environmental variables.
Direct evidence of the ability of C. virginica to filter, ingest, and assimilate marine bacteria has been published (Crosby et al. 1990).Detritus and bacteria are food sources for C. virginica, although phytoplankton, nanozooplankton, non-cellulosic particulates, and dissolved organic matter are also required (Landon and Newell 1990).In the present investigation, Chl a was significantly higher in RL than in CL.However, no direct relationship was established between oyster growth and mortality rates with the Chl a concentration in either site.Moreover, no significant differences were detected in final oyster height and daily growth rates between the lagoons at the end of the nursery stage.Thus, it is likely that the difference in food availability associated with Chl a in both sites was compensated by the availability of other food sources.Piyathilaka et al. (2012) found a positive relationship between the growth rate of Crassostrea madrasensis and total dissolved solids.The authors argued that dissolved solids were indicative of food abundance, as oysters feed on phytoplankton, detritus from sediment, bacteria, and dead phytoplankton and zooplankton.We found no evidence of dissolved solids influencing oyster growth.However, a possible positive effect on survival was observed in CL, where higher dissolved solids were associated with better survival.However, this result must be taken cautiously, as a positive correlation may exist between salinity and total dissolved   solids (Boyd 2020, Adjovu et al. 2023).Thus, total dissolved solids may appear to affect survival merely because this variable is correlated with salinity, with the latter being the actual cause influencing survival.To summarize, low salinity, low dissolved oxygen, and possibly low food availability in CL could increase oyster mortality, particularly at the end of the nursery stage.
In RL, faster growth rates were associated with high temperature and pH values.Temperature increases within acceptable limits promote oyster growth (Lowe et al. 2017), and seawater acidification appears to affect calcifying marine organisms (Clements et al. 2021).Beniash et al. (2010) showed that pH values of ~7.50 inhibited the shell and soft body growth of C. virginica when compared with control organisms growing under pH conditions of ~8.20.Thus, the inferior growth results obtained in RL at low pH may be due to an adverse effect on calcification when conditions in the lagoon were more acidic.
According to Rybovich et al. (2016), C. virginica can survive extended periods of low salinity (<5 ppt) at low  The authors also noted that the combination of high water temperature (>30 °C) and high salinity (>15 ppt) resulted in even higher oyster mortality.In contrast to those results, salinity and temperature in RL were high (31.02°C and 35.02 ppt), but oyster mortality was negligible (lower than 1.0%), and survival was higher than in CL, where salinity was generally lower.Thus, even when the conditions are not those commonly reported as suitable for C. virginica production, high environmental stability (mainly in salinity, as seen in RL) may compensate for suboptimal rearing conditions.
Recent studies have shown that the hypoxia and salinity tolerance of C. virginica varies among populations in the Gulf of Mexico.For instance, Coxe et al. (2023) showed that distinct populations might be better adapted to tolerate extended periods of hypoxia during summer.Marshall et al. (2021) tested the salinity tolerance of oyster progeny from wild broodstock collected from estuaries with different mean salinities (7.40-35.50 ppt).They found that the progeny were better able to tolerate the salinity levels of the estuary where the corresponding broodstock was collected.
The spat used for this study was produced from a cultivated broodstock maintained in an estuary with a mean salinity of 18 ppt; the spat was then acclimated to a salinity of 24 ppt in the hatchery.The potentially high plasticity of the broodstock may have contributed to the excellent performance observed at high salinities and temperatures in this study, particularly in RL, although further studies are required to evaluate this possibility.
The modeling approach used in the present study helped us analyze and understand the dynamics of oyster nursery production and its relationships with the environmental variables of notably different lagoons.The growth and survival nonlinear regression models, namely Eq. (1), Eq. ( 2), and Eq. ( 3), proved to be statistically adequate for describing the shell height and survival of the oysters grown in both lagoons.It is    worth noting that the model used to describe oyster survival in CL assumed that the mortality rate increased over time, thus accounting for the worsening environmental conditions in the lagoon as the nursery phase progressed.

Conclusions
The present study has shown that nursey production of C. virginica is technically feasible in the tropical coastal lagoons of CL and RL in Mexico.The results indicate that even when the environmental conditions are not optimal, rearing oysters under relatively stable conditions (RL) may be preferable to rearing oysters under more unstable conditions (CL).Further research should focus on the tolerance of C. virginica to the stressful conditions associated with genetic and physiological characteristics of the oyster populations in the southern Gulf of Mexico and the Caribbean Sea.The economic feasibility of nursery production in CL and RL should also be investigated.

Figure 1 .
Figure 1.Study sites in the state of Yucatán, Mexico.Oysters were reared in the tropical coastal lagoons of Celestún (CL) and Rio Lagartos (RL).Cultivation sites are indicated with marker pins ( ).

Figure 2 .
Figure 2. Growth models fitted using Eq.(1) and Eq.(2) to the mean height data of Crassostrea virginica reared off-bottom during the nursery stage in the coastal lagoons of Celestún (CL) and Rio Lagartos (RL), respectively.The vertical bars indicate standard error (height).

Figure 3 .
Figure 3. Survival curve, Eq. (3), fitted to survival percentage data during the nursery stage of Crassostrea virginica reared in an offbottom system in the coastal lagoon of Celestún (CL).No estimates of the standard error were available because survival values corresponded to single counts.

Figure 4 .
Figure 4. Temperature (T), salinity (S), dissolved oxygen (DO), pH, chlorophyll a (Chl a) concentration, and total dissolved solids (TDS) during the nursery stage of Crassostrea virginica reared in an off-bottom system in the coastal lagoon of Celestún (CL).

Figure 5 .
Figure 5. Temperature (T), salinity (S), dissolved oxygen (DO), pH, chlorophyll a (Chl a) concentration, and total dissolved solids (TDS) during the nursery stage of Crassostrea virginica reared in an off-bottom system in the coastal lagoon of Rio Lagartos (RL).

Figure 7 .
Figure 7. Relationship between the coefficients of variation of oyster height and salinity during the nursery stage of Crassostrea virginica reared in an off-bottom system in the coastal lagoon of Celestún (CL).Roman numerals indicate the periods when oyster growth increased (I and III) or stabilized (II and IV), which were associated with stable or unstable (decreasing) salinity values, respectively.

Figure 6 .
Figure 6.Periods when oyster growth increased (I and III) or stabilized (II and IV), which were associated with stable or unstable (decreasing) salinity values, respectively, during the nursery stage of Crassostrea virginica reared in an off-bottom system in the coastal lagoon of Celestún (CL).

Table 1 .
Fitting performance of equations (1) and (2) to oyster growth data by study site.Lower values of the Akaike Information Criterion (AIC) indicate better performance.

Table 2 .
Regression analysis results using Eq.(1) and Eq.(2) for growth (height) and survival of Crassostrea virginica reared offbottom during the nursery stage in the Celestún and Rio Lagartos lagoons, Mexico.SE = standard error.

Table 3 .
Regression analysis results using Eq.(3) for survival of Crassostrea virginica reared off-bottom during the nursery stage in the Celestún Lagoon, Mexico.SE = standard error.

Table 4 .
Linear regression equations relating mortality rate (MR) and growth rate (GR) to environmental variables.S = salinity; DO = dissolved oxygen; TDS = total dissolved solids.