The genus Centropomus has a wide distribution throughout the American continent: from the west coast of the Atlantic Ocean, starting from North America (Florida) to South America (Rio de Janeiro, Brasil); including the Gulf of Mexico in the Caribbean Sea (1). The common snook Centropomus undecimalis is a migratory, euryhaline and hermaphrodite fish (2).

This species is considered a promising candidate for aquaculture due to its adaptation to abrupt environmental changes such as salinity, oxygen, adaptation to captivity and commercial diets; moreover, it has a good quality flesh which is widely marketed (3). The fisheries of common snook represent one of the main economical and subsistence activities performed by coastal communities, which in turn has caused the overexploitation of the common snook natural stocks (2).

The aquaculture field represents an interesting solution to the decrease in the availability of the resource. However, there are several factors that must be investigated in order to establish the viability of the common snook culture; among those factors is feeding, which represents a crucial aspect due to the nutrition specificity, its acceptance, availability and economical costs (4).

Therefore, the present study evaluated the performance of common shook juveniles when fed with three experimental diets which are commercially available in the country: trout (45% protein) and cobia (50% protein) feed and raw fish (17.5% protein). The aim of the study is to contribute baseline knowledge for the construction of a technological protocol of the culture of common snook, which in turn could increase the socieconomical alternatives for local communities and the restocking programs.

Statystical Analysis. The obtained data was analyzed for normality and variance homogeneity by means of Kolmogorov-Smirnov and Bartlett tests respectively. After normality and homogeneity were fulfilled, a one-way variance analysis (ANOVA) was performed in order to establish if significantly differences among the groups existed (p=0.05). When this was the case, a Tukey HSD post-hoc test was applied in order to examine the diet effect among the treatments. Otherwise, the test Mood median test was performed to evaluated the hypothesis of similarity among medians trough Chi-squared (X2) so the best diet for common snook C. undecimalis could be identified. Statistical Analysis were applied using the Statgraphic Centurion XVI.I software (6).

Location and study area description. This study was performed at the Experimental Aquaculture Station located at University del Magdalena (Santa Marta, Colombia), with geographical coordinates 11°13’24.82” of Northern latitude and 74°11’07.14” of West Longitude. The water used for the study was taken from a water bore and no further treatment was applied.

Biological Material and experimental units. 108 common snook Centropomus undecimalis juveniles were used (282.8±11.5 g weight and 34.2±5.5 cm length). Fish were captured with an 8 mm net at the Ciénaga Grande de Santa Marta, and were acclimated in a 1 m3 tank for 48 h during which salinity was decreased from 33.9±3.3 UPS to 0-0.5 UPS. Once acclimation ended, fish were treated with a prophylactic protocol involving formalin baths (150 ppm) for 15 min, and then fish were transferred to a 620 m3 water pond divided in nine corrals.

Experimental diet. Three treatments of commercial feeds were applied: D1 (trout fed – 45% protein, 12% humidity, 14% fat, 12% ash and 6% fiber), D2 (cobia fed – 50% protein, 12% humidity, 15% ash, 2.5% fiber and 8% fat) and D3 (raw fish – 17.5% raw protein, 1.75% ethereal extract, 2.89% ash and 2.22% carbohydrates). Each treatment had 36 fish and three replicates randomly distributed. Feed was provided to the fish twice a day (11:30 y 17:00 h) at 3% of biomass for 120 days.

Every 20 days biometric data was recorded (total length and wet weight) in 24 fish per treatment, which were randomly collected. Weight gain (g) was recorded with a digital scale (LEXUS MIX-SL, precision=0.01 g) and length increase in cm. In order to avoid stress during the sampling fish were sedated with clove oil (20-40 ppm) and wet clothes were used for manipulation to prevent scales and mucus loss.

Evaluated parameters. As indicators of performance the following variables were evaluated: length-weight relationship, feed conversion factor (FCF), condition factor (CF), specific growth rate (SGR) and final survival.

Length-weight relationship: evaluated through lineal regression (least squares), calculating the values for a and b from the equation:

W=aLb [1]

W being the total weight in g and L the total length in cm; a and b are constants obtained from the potential regression analysis (5).

Feed conversion factor (FCF). This indicates the amount of feed that is needed in order to increase a weight unit (2).

FCA= food ingested/ (final weight – initial weight) [2]

Condition Factor (K). Is the final condition state of an organism, calculated with Fulton’s index:

K=100 (W/L3) [3]

W being the wet weight in g and L the length in cm (2).

Specific growth rate (SGR). SGR estimates the growth daily rate:

SGR=100*[LnPf – LnPi]/t [4]

LnPf= natural logarithm of the final weight, LnPi= natural logarithm of the initial weight y t being the time in days (2).

Survival was registered at the end of the experimental period.

Every day, physical and chemical parameters were recorded: temperature (°C±0.2), salinity (UPS ±1 ), dissolved oxygen (mg/L ±0.5 ) and pH (±0.02) by means of a multiparameter (WTW Multi 350i); a final nutrients follow up was also performed: alkalinity (mg/L), nitrates (mg/L), nitrites (mg/L), hardness (mg/L), phosphates (mg/L) y chlorine (mg/L) by means of colorimetric kits (Merck KGaA).

Statystical Analysis. The obtained data was analyzed for normality and variance homogeneity by means of Kolmogorov-Smirnov and Bartlett tests respectively. After normality and homogeneity were fulfilled, a one-way variance analysis (ANOVA) was performed in order to establish if significantly differences among the groups existed (p=0.05). When this was the case, a Tukey HSD post-hoc test was applied in order to examine the diet effect among the treatments. Otherwise, the test Mood median test was performed to evaluated the hypothesis of similarity among medians trough Chi-squared (X2) so the best diet for common snook C. undecimalis could be identified. Statistical Analysis were applied using the Statgraphic Centurion XVI.I software (6).

Monthly mean values of physical and chemical parameters did not differ among treatments or experimental units (Table 1). Water quality indexes showed that nitrites, nitrates and chlorine values were among the recommended range for fish culture, while alkalinity (>150 mg/L CaCO3), hardness (>300 mg/L CaCO3) and phosphates (minor eutrophication >0.2 mg/L PO4 –P), were high (Table 2).

Table 1 Table 1 Table 1. Physico-chemical parameters during the bioassay juvenile C. undecimalis grown in a freshwater pond.

Parameter	Month 1	Month 2	Month 3	Month 4
Temperature (°C)	30.5±0.63	30.9±0.42	29.8±0.33	29.7±0.24
O2 (mg/L)	4.75±1.05	4.66±0.21	4.53±0.23	4.62±0.35
pH	6.94±0.60	7.64±0.11	7.55±0.12	7.68±0.20
Salinity (UPS)	0.28±0.04	0.13±0.02	0.16±0.02	0.23±0.02

Table 2 Table 2 Table 2. Water nutrients end of the experimental cultivation of juvenile sea bass in fresh water pond.

Nutrients	Value
Alkalinity (mg/L CaCO3)	283±5.2
Nitrites (mg/L)	600±13.0
Nitrates (mg/L)	2.34±1.31
Dourness (mg/L CaCO3)	0.32±0.3
Phosphates (mg/L PO4 –P)	0.47±0.1
Chlorides (mg/L)	155±4.0

In Table 3 the growth indexes K and SGR are shown to be higher under D2, whose values significantly differed to the rest of the diets. The FCA of D2 did not differ with D3 data but both values differed with D1 FCA. Furthermore, the Mood Median test indicated statistical differences in weight gain between D2 when compared to D1 and D3 (p<0.0428). Fish fed with D2 recorded the highest final weight (630.2±29.6 g), SGR, (0.65±0.09 g/day), and K (0.81±0.06).

Tabla 3 Tabla 3 Tabla 3. Factors related to the growth of juvenile common snook kept in ponds and fed fresh water with three commercial diets (D1-truchina, D2-cobia, D3-fresh fish) for 120 days.

Tratamiento.	D1	D2	D3
Pi (g)	286.8±10.5	286.7±3.80	276.0±11.7
Pf (g)	436.1±82.9a	630.2±29.6b	472.0±21.7a
Li (cm)	34.3±0.31	34.2±0.52	34.1±0.47
Lf (cm)	38.8±2.50	42.1±0,53	39.0±0,81
S%	51.4	50	77.8
FCA	12.0a	3.76b	3.05b
K	0.73±0.04a	0.81±0.06b	0.76±0.05a
SGR (g/día)	0.31±0.33a	0.65±0.09b	0.44±0.13a
Equal exponents do not differ statistically (p>0.05). Trat.= Tratamiento, Pi=Peso inicial, Pf=Peso final, Li=Longitud inicial Lf=Longitud final, S=supervivencia, FCA=Factor de conversión alimenticia, K=Factor de condición, SGR=Tasa específica de crecimiento.

The length-weight relationship showed isometric growth, with an estimated relationship (W=0.003L3.253) being b=3.2, with 97% of weight variability explained by the model.

Overall, from the 100 and 120 days of the experimental period, D2 registered the highest values, which are supported by the statistical analysis (p<0.05) (Figure 1).

Figure 1 Figure 1 Figure 1

No statistical differences were detected between the survivals of the treatments (p>0.05). The survivals for D1=51.4%, D2=50% and D3=77.8% are considered acceptable for the three treatments (Figure 2).

Figure 2 Figure 2 Figure 2

During the study period, the physical and chemical parameters registered acceptable values for aquaculture purposes. Regarding the hardness and alkalinity data, which were above the suggested range, common snook seemed to have acclimated accordingly. Although the high levels of phosphates produced a general eutrophication in the culture system, this condition did not affect any treatment or experimental unit in particular.

It is not common for snook to be cultured under fresh water conditions, however the results obtained in the present study are similar to the highest data previously reported in a similar study of juvenile’s nutrition under sea water conditions (7), suggesting that water salinity does not affect this species’ growth - at least at this developmental stage. In an earlier study of common snook aquaculture, an experimental feeding based on live tilapia (Oreochromis niloticus) obtained acceptable growth results (0.50 g/day), which imply that this species has an interesting potential to be cultured (3). In the present study, the daily weight gain was of 0.65 g/day, which shows that growth can be improved by means of high quality feeding.

In this study, the highest data for final growth were obtained with the cobia feed (D2), which not only had the highest protein content (50%) but also all the necessary requirements for optimal performance of Rachycentron canadum, a marine fish. In a similar study (8), fed semi-humid diets with a variable percentage of protein to common snook, obtained as a result that diets with a ≥40% protein brought out the highest growth performances in terms of final weight and food conversion factor. This last result is similar to the one obtained in the present study (8).

Treatment D1 (trout feed) brought out the lowest FCF and SGR values which contrast with the results found by Gracia-López et al (8), who registered the second highest records of these two variables, plus final growth under the trout diet. Due to the lack of information regarding the specific formulation of commercial diets, it is difficult to explain the results obtained in the present study and why they differed from previous studies. However, regarding the results reported by Gracia-López et al (8), their study was performed with smaller fish. Moreover, we experienced a delay in the acceptance of the trout diet, which might have been due to the size of the pellets.

Treatment D3 (raw fish) did not show significant differences with D2 regarding FCF, although differences were found in K and final weight data between these two treatments. Interestingly, the results obtained in the present study contrast sharply with those registered in Reyes et al. (2004) research, as they reported a lower performance in terms of FCF and growth of similar size juveniles of common snook that were fed with a diet composed of fish meal and raw fish (O. oglinum y Gerres sp) (9).

The above mentioned results show that diet formulation for fish is a complex process that it has not been well established for many species and even in those cases where a commercial specific diet is commercially available, it still might bring about metabolic problems that negatively affect its growth and survival (10). Moreover, artificial feeds designed for fresh water species should not be used in estuarine and marine fish due to the wide physiological differences existing among them (i.e. fat acid requirements) (9).

The results obtained in other studies (11,12) regarding length-weight relationship also suggest an isometric growth as values for this parameter were between b= 2.5 – 3.5, indicating that the common snook increases proportionally in length and weight.

In general, the survival data recorded by the present study are acceptable for fish that have to undergo captivity acclimation. However, previous studies have reported 90% survivals (3, 4, 13). For this particular issue, no causes could be singled out as mortalities were present evenly in all treatments.

As conclusions, the culture of common snook under fresh water conditions fed with cobia commercial feed is a viable alternative. However, more studies are needed in order to design a specific feed for the specie. Research that studies the performance on common snook under low-cost and high-availability feeds are needed with the purpose of developing an aquaculture technical protocol that could be transferred to artisanal farmers.