LARVICULTURE OF SIAMESE FIGHTING FISH Betta splendens IN LOW-SALINITY WATER

1Universidade do Estado de Santa Catarina, Avenida Luis de Camões, 2090, 88520-000, Lages, SC, Brasil, thiagofabregat@hotmail. com (autor correspondente) 2Fundação Instituto de Pesca do Estado do Rio de Janeiro, Laboratório de Reprodução de Peixes Nativos e Exóticos. Cordeiro, RJ, Brazil. Cep 28540-000. 3Universidade Federal de Minas Gerais, Escola de Veterinária, Laboratório de Aquacultura, CEP 567, 30123-970, Belo Horizonte, MG, Brazil. 4Universidade Estadual Paulista, Centro de Aquicultura, 14884-900 Jaboticabal, SP, Brazil. ABSTRACT


INTRODUCTION
Ornamental fish is a multimillionaire industry around the world (CHAPMAN et al., 1997;ANJOS et al., 2009).Nevertheless, significant number of fish is still arising from exploitation on natural stocks (TLUSTY, 2002;PELICICE and AGOSTINHO, 2005;RAGHAVAN et al., 2013;SAMPAIO AND OSTRENSKY, 2013).One alternative to decrease the capture from the natural environment is the production of ornamental species in captivity (TLUSTY, 2002;BARBIERI et al., 2014).
Several studies have been developed to improve the knowledge on the biology and management during the rearing of ornamental species, such as: feeding management (ZUANON et al., 2004), techniques and system of production (NG et al., 1992;HALACHMI, 2006;ZUANON et al., 2011), nutrition (THONGPRAJUKAEW et al., 2011;ZUANON et al., 2011), reproduction (ROY andYANONG, 1996), behavior (HALLER and WITTENBERGER, 1987), among others.However, the number of species for ornamental purposes is high, and there is lack of information about larviculture even for those produced globally on a large scale.
The use of sodium chloride (NaCl) in water has proved to be an important rearing protocol during the production of larvae (SANTOS and LUZ, 2009;LUZ and SANTOS, 2010;LUZ et al., 2012;JOM ORI et al., 2012;JOMORI et al., 2013) and juveniles (ZUANON et al., 2009;SALARO et al., 2012) of freshwater fish species.In literature, tolerance of Betta splendens to salinity was been reported by ZUANON et al. (2009) and PUELLO-CRUZ et al. (2010).However, there are no references of the median lethal salinity estimates to Siamese fighting fish larvae.Furthermore, information on the use of Artemia nauplii at first feeding in low saline water is missing for this species.In fact, the low-salinity water does not affect survival of several freshwater fish larvae and improves live food utilization, mainly when larvae are fed with Artemia nauplii (JOMORI et al., 2013), because the saline water extends the lifetime of this microcrustacean (JOMORI et al., 2012), allowing longer exposition to the larvae and improved availability.
B. splendens is a small native fish from Asia, Mekong basin, but was spread world widely.The species is one of the most important ornamental fish in United States of America (CHAPMAN et al., 1997).However, information about the handling during its rearing is still scarce, mainly in larviculture, which is considered a bottleneck in fish production (PORTELLA and DABROWSKI, 2008).We hypothesize here that the use of slightly saline water can improve the larviculture of Siamese fighting fish B. splendens, optimizing the use of Artemia nauplii when higher densities of prey are used.Therefore, this study aimed to determine the salinity tolerance of Siamese fighting fish larvae, and also to improve the protocol of larviculture, by evaluating the larvae rearing in low-salinity water with different prey densities.

METHODS
The experiment was approved by the Ethics Committee of the Santa Catarina State University (protocol 4774111115).B. splendens larvae were obtained from natural reproduction of broodstock kept in the laboratory at 28°C.At the end of yolk-sac phase (three days post-hatching), the larvae were submitted to two experiments.In the Experiment 1, the larvae were exposed to different concentrations of NaCl in water to access the median lethal concentration (LC 50 ).In the second experiment, we evaluated the effects of low-salinity and prey density on the larviculture.

Experiment 1
Four hundred larvae of B. splendens (3.07 ± 0.20 mm and 0.17 ± 0.05 mg) were stocked in 40 aquariums of 250 mL, totaling 10 larvae per aquarium.The aquariums were displayed in a water bath kept at 28.4 ± 0.7°C using electric thermostat, and were supplied with low artificial aeration.The experiment was carried out in a completely randomized design, with 10 treatments (saline water) with four replicates.The commercial salt (iodine-free NaCl) was used to prepare the different saline waters.
Data of cumulative mortality were used to estimate median lethal concentration (LC 50 ) of NaCl in water and their respective confidence intervals (95%) for safe salinity concentration.

Experiment 2
Three hundred and sixty larvae of B. splendens (3.12 ± 0.21 mm and 0.25 ± 0.04 mg) were distributed in 24 1 L aquariums (15 larvae aquarium -1 ) supplied with low artificial aeration.The temperature was maintained constant at 27.4 ± 1.7°C using water bath, and the other water quality variables were: dissolved oxygen 6.0 ± 0.3 mg L -1 , pH 7.65 ± 0.24 and total ammonia 0.09 ± 0.05 mg L -1 .
The study was carried out during 15 days in factorial design 3 x 2, with three low-salinity concentrations (0, 2 and 4 g NaCl L -1 ) and two increasing levels of prey densities (N 50 and N 100 ), with four replicates.Larvae were fed with freshly hatched Artemia nauplii (24-h post cyst incubation) and the live food amount was divided in three meals a day, at 8 am, 12pm and 4pm.At the beginning of exogenous feeding, larvae were fed with Artemia nauplii at daily density of 50 and 100 nauplii larvae -1 from the first to the third day in N 50 and N 100 levels, respectively.The prey densities increased to 100 and 200 nauplii larvae -1 per day from the fourth to sixth day, 150 and 300 nauplii larvae -1 per day from the seventh to ninth day, 200 and 400 nauplii larvae -1 from the tenth to twelfth day and 250 and 500 nauplii larvae -1 from the thirteenth to fifteenth day, in N 50 and N 100 levels, respectively.During the experiment, after the last meal, the residues accumulated on the bottom and dead larvae were removed by siphoning, and approximately 90% of the volume was renewed with clean water.
The total length and weigh of the B. splendens larvae were evaluated using a digital calliper (Starret ® ) and an electronic scale (Marte ® ), respectively.The growth data were used to calculate the weight gain (WG = final weight -initial weight), final biomass (B = final number of larvae x final weight) and specific growth rate [SGR = (ln final mean weight -ln initial mean weight) 100/ time interval (in days)].The survival at the end of Experiment 2 was expressed in percentage.

Statistical analysis
Statistical analyses were performed using the software Statistical Analysis System -SAS Institute, version 8.0.Percent results (survival rates and SGR) were arcsine transformed before being analyzed but only the original data are presented.Tests were performed to verify error normality (Cramervon Mises) and homoscedasticity of the variance (Levene).In experiment 1 comparison among survival rates were made by one-way ANOVA followed by the Tukey's Test.In addition, the median lethal concentration (LC 50 ) of NaCl in water and their respective confidence intervals (95%) for safe salinity level, were estimated using the software Trimmed Spearman Karber (HAMILTON et al., 1977).The data from Experiment 2 were compared by parametric Factorial ANOVA (salinity x prey density), and means were compared using Tukey's multiple range tests, at 5% probability level.

Experiment 1
Total mortality was observed in salinities above 10 g NaCl L -1 after 24 h of exposure.The salinity of 10 g NaCl L -1 provided decreasing survival (P<0.05) after 24 and 48 hours of exposure.After 72 and 96 hours, larvae exposed to salinity of 10 g NaCl L -1 showed the lower survival (P<0.05) and the salinities of 0, 2 and 4 g NaCl L -1 resulted in higher survival, with no statistic differences among them.Salinities of 6 and 8 g NaCl L -1 showed intermediate survival rates (P<0.05)(Table 1).
The median lethal salinity concentrations (LC 50 ) for B. splendens larvae were estimated after 24, 48, 72 and 96 h of exposure, as well as the respective safe salinities.The results demonstrate that toxicity of NaCl increased over time.The LC 50 for B. splendens larvae changed from 11.7 at 24 h to 7.1 g NaCl L -1 at 96 h (Table 2).

Experiment 2
Larvae of B. splendens reared at different salinities and prey densities showed distinct growth parameters.The mean body weight, total length, final biomass and SGR showed interactions (P<0.05) between salinity levels and prey densities.On the other hand, survival did not show interaction between the factors (Table 3).
At the higher prey density (N 100 ), the mean body weight, total length (Table 4), final biomass and SGR (Table 5) were, in general, higher (P<0.05).When analyzing the saline influence, we observed that the lower prey density (N 50 ) resulted in similar mean values of all parameters (P>0.05),except total length in freshwater (Table 4).
However, the higher level of prey density (N 100 ) resulted in higher (P<0.05)weight and length to B. splendens larvae reared at salinities of 2 and 4 g NaCl L -1 than in freshwater (Table 4).
Final biomass and SGR were higher (P<0.05) in the N 100 treatments, independently of the salinity used.However, in N 50 treatments, there were no statistic differences in final biomass and SGR of B. splendens larvae reared at the different salinities.hand, the larvae at N 100 showed higher (P<0.05)average values when reared at salinities of 2 and 4 g NaCl L -1 compared to those reared in freshwater (0 g NaCl L -1 ).The B. splendens larvae survival did not differ amongst the treatments (Table 6).Table 6.Average values (± standard deviation) of Betta splendens survival rates after 15 days of exogenous feeding.

DISCUSION
The larvae of the ornamental fish B. splendens showed susceptibility to high-salinity water.However, the intensive larviculture of this species could be carried out successfully in low-salinity conditions.In such case, a series of factors may be improved during the larviculture, as the physiological condition by decreasing the energetic requirements for ionic and osmotic regulation (ALTINOK and GRIZZLE, 2001), the prevention of diseases (GARCIA et al., 2007) and the increased lifetime and availability of the live food (Artemia nauplii) (JOMORI et al., 2012).
The median lethal salinity (LC 50 ) for B. splendens larvae changed during the time, from 11.7 at 24 h to 7.1 g NaCl L -1 at 96 h.The longer the exposure time, more larvae were dying.A possible explanation for the increased toxicity over time may be that mortality occurred as the energy reserves of the larvae were depleted by the energy expenditure of the osmoregulation.For instance, salinity of 10 g NaCl L -1 killed only 12.5% of B. splendens larvae during the first 24 hours, but after 96 hours it ended up being lethal for 97.5%.However, the LC 50 96 h for B. splendens larvae was lower than the reported for adult of the same species, 11.88 g NaCl L -1 (ZUANON et al., 2009).The saline tolerance may change during fish development, as already showed for bonefish (Albulidae: Albula) (PFEILER, 1981) and mangrove red snapper (Lutjanus argentimaculatus) (ESTUDILLO et al., 2000).
The differences observed in median lethal salinity between the larval and adult stages may be related to the ontogeny of osmoregulation in postembryonic fish, since during the larval phase the osmoregulatory function shifts from the skin to the gills, which become the main osmoregulatory site (VARSAMOS et al., 2005).According to the authors, the main difference is due to Na + /K + -ATPase activity, which varies ontogenetically in the tegument and gill.Future studies should be developed to characterize the Na + /K + -ATPase activity during the initial development of stenohaline fish species.
The rearing of B. splendens larvae in low-salinity conditions resulted in better or similar growth and survival rates compared to freshwater rearing conditions, as previously observed for other stenohaline species (BORODE et al., 2002;SANTOS and LUZ, 2009;LUZ et al., 2012;JOMORI et al., 2012;2013).For instance, the growth of Colossoma macropomum, Astronotus ocellatus, Brycon amazonicus and Leporinus macrocephalus larvae was positively affected when they were reared in low-salinity water (JOMORI et al., 2013).The growth and survival of A. ocellatus and B. amazonicus were improved in freshwater and at 2 g NaCl L -1 in the water.The species C. macropomum and L. macrocephalus grew better in water at 2 and 4 g NaCl L -1 ; however, the survival of C. macropomum was lower at 4 g NaCl L -1 whereas the survival of L. macrocephalus larvae was not affected.Taken together, it is possible to say that B. splendens larvae follow the tendency observed in larvae of other stenohaline fish species.
Artemia nauplii has been used successfully as live feed during the intensive larviculture of freshwater species (PORTELLA and DABROWSKI, 2008;LUZ and SANTOS, 2010;JOMORI et al., 2012;PORTELLA et al., 2014), and also in Siamese fighting fish larval rearing (OGATA and KUROKURA, 2012;FOSSE et al., 2013).However, the lifetime of the Artemia nauplii in freshwater is reduced, and the use of low-salinity water may improve its lifetime (JOMORI et al., 2012), providing longer period of live prey exposure to the fish larvae.In this situation, the higher prey density in low-salinity water may favor the possibility of larvae feeding, by extending the period of feeding activity and, by having a higher amount of live prey available, resulting in higher growth rate.In the present study, this was evidenced by the interaction between the factors salinity and density of prey.The higher density of prey resulted in higher weight and length of B. splandens larvae reared at salinities of 2 and 4 g NaCl L -1 than in freshwater.Similar situation was observed during the intensive larviculture of other stenohaline fish species, such as Prochilodus costatus, Lophiosilurus alexandri (SANTOS and LUZ, 2009), Astronotus ocellatus, Brycon amazonicus and Leporinus macrocephalus (JOMORI et al., 2013).This is the first report on the use of common salt to improve the Artemia nauplii utilization in the larviculture of B. splendens.
Our findings confirmed that low-saline water is a safe protocol and can be used during B. splendens larviculture.In addition, the results of the interaction between the factors NaCl concentration and prey density show that the longer availability of Artemia nauplii improved the growth performance of B. splendens larvae.

CONCLUSION
The CL 50 for B. splendens larvae ranges from 11.7 to 7.1 g NaCl L -1 during the first 96 h of observation.The use of low-saline water does not affect the survival of B. splendens larvae and optimizes the use of Artemia nauplii when higher prey concentrations are used.
To improve the larviculture of B. splendens we can recommend the use of low-salinity water (2 to 4 g NaCl L -1 ) and an initial prey density of 100 nauplii larvae -1 , followed by periodic increasing amounts during the first 15 days of exogenous feeding.

Table 1 .
Survival (%) of Betta splendens larvae exposed to different salinities for 96 hours.

Table 2 .
Median lethal salinity concentrations (LC 50 ) and safe salinities (confidence interval) throughout the 96 h test with Betta splendens larvae.

Table 3 .
F and average values of weight, total length (TL), final biomass, specific growth rate (SGR), and survival during the larviculture of Betta splendens at different salinities and prey densities.
AaMeans followed by the same letters (A, B in the rows and a, b, in the columns) for each parameter did not differ by Tukey's test (p<0.05)N 50 and N 100 : Larvae fed initially with 50 and 100 nauplii larvae -1 per day, respectively.

Table 4 .
Interaction (Salinity × Prey density) average values (± standard deviation) of weight and total length after fifteen days of active feeding for Betta splendens larvae.Means followed by the same letters (A, B in the rows and a, b, in the columns) for each parameter did not differ by Tukey's test (p<0.05)N 50 and N 100 : Larvae fed initially with 50 and 100 nauplii larvae -1 per day, respectively.

Table 5 .
Interaction (Salinity × Prey density) average values (±standard deviation) of final biomass specific growth rate (SGR) after fifteen days of active feeding for Betta splendens larvae.