Cannibalism , growth performance , and body composition of giant trahira juveniles under different photoperiods

The objective of this work was to determine the effect of photoperiod on the cannibalism, growth, and body composition of giant trahira (Hoplias intermedius) juveniles. A completely randomized design was employed, with four photoperiods – 0L:24D, 6L:18D, 12L:12D, and 18L:6D, hours of light (L): dark (D) – and four replicates, for a period of 90 days. The analyzed parameters were: cannibalism, growth in weight and length, food consumption, feed conversion, protein efficiency rate, muscle growth, and body composition (moisture, lipid, ash, and crude protein). Fish from the 12L:12D photoperiod showed the highest survival rate and no cannibalism, as well as a great uniformity for length and mass values at the end of the experiment. The best photoperiod for giant trahira juveniles is 12L:12D, as these fish showed growth uniformity and no cannibalism, while fish from the 6L:18D and 18L:6D groups had a greater diameter of white muscle fibers and a higher frequency of greater diameter fibers. Giant trahira showed the lowest value for lipids at 6L:18D and the highest value at 12L:12D. Photoperiod affects survival and cannibalism; regarding body composition, only lipids are influenced by photoperiod.


Introduction
The giant trahira Hoplias intermedius (Günther, 1864) has aroused the interest of researchers and farmers as a potential fish for aquaculture purposes.This fish adapts well to captive conditions, has quality meat, high rate of mass gain, and shows features for sport fishing.However, this species has carnivorous habits, which may result in low survival rate, making it difficult to cultivate it (Luz et al., 2000).Pesq. agropec. bras., Brasília, v.53, n.6, p.664-672, June 2018 DOI: 10.1590/S0100-204X2018000600002 Several factors have been linked to cannibalism.Population factors, as size and density, have been correlated with the incidence of cannibalistic behavior (Naumowicz et al., 2017).Besides, environmental factors such as feed, diet supplementation, refuges, color, and shape of tanks, light intensity, photoperiod, and temperature can also increase cannibalism (Naumowicz et al., 2017).However, the size uniformity of H. intermedius has been considered important to minimize cannibalism (Luz et al., 2000).
The relationship between cannibalism and photoperiod has been shown in some species.In Steindachneridion scriptum larvae, the absence of light (0L:24D photoperiod) provided better effect on growth, but high mortality and cannibalism were observed.Dark conditions provided the chance of cannibalism and agonistic interactions (Schütz et al., 2008).Nevertheless, in Pseudoplatystoma corruscans larvae, the cannibalism was lower with decreasing photoperiod (Campagnolo & Nuñer, 2008), showing that there are different responses to photoperiod for each species of fish.
Changes in animal growth can be caused by photoperiod, as it acts on the somatotrophic axis, mass gain, and food intake, in some fish species (Taylor & Migaud, 2009;Biswas et al., 2010;Falcón et al., 2010).Biswas et al. (2008Biswas et al. ( , 2010) ) showed that the growth of Pagrus major, subjected to long photoperiods, was stimulated with a significant increase of food intake.
The postembryonic growth of muscle involves myogenic progenitor cells through the process of hypertrophy and hyperplasia (Johnston, 2006).Hypertrophy is characterized by the increase in length and diameter of the existing muscle fibers, while hyperplasia is the formation of new fibers (Johnston et al., 2009).In adults and juveniles, abiotic factors such as photoperiods affect the signaling pathways that regulates the proliferation and differentiation of myogenic progenitor cells.The photoperiod can control the myoblast proliferation, which results in proportional increases of the myotube production and the myonuclear content of fibers (Johnston, 2006).Therefore, the photic environment parameter can determine myogenesis rate, and can also change the distribution of muscle fibers in both number and size.
Additionally, the photoperiod can cause changes in the body composition of fish, as it influences the stimulation of the metabolic activities, feeding strategy and the efficiency of the nutrient use of some fish (Reynalte-Tataje et al., 2002;Veras et al., 2014).Veras et al. (2014) found that the 18L:6D photoperiod contributes to the concentration increase of crude protein, and less fat deposition in the fillet of Oreochromis niloticus, since it influences the metabolic rates for the mobilization of nutrients, particularly of protein deposition and utilization of body fat.
All this information shows how much the photoperiod can interfere in the behavior, development, and physiology of fish.The cultivation of H. intermedius is difficult due to this fish carnivorous habits, and photoperiod is closely related to this factor, acting directly or indirectly on fish behavior.
The objective of this work was to evaluate how different photoperiods affect the cannibalism, growth, and body composition of H. intermedius juveniles, since there is still little information available on these issues for this species.

Materials and Methods
The study was conducted at the photoperiod laboratory of fish in the Biotério Central of Universidade Federal de Lavras, in the municipality of Lavras, in the state of Minas Gerais (MG), Brazil, from June 02 to October 28, 2012.
Juveniles of H. intermedius with 0.84±0.36g and 3.5±0.2cm, at 30 days after hatching of the same spawning, were used.The animals were purchased from the fish farm station of Furnas Centrais Elétricas S.A., located in the municipality of São José da Barra, MG, Brazil.Fish were stocked at a density of 3 juveniles L -1 in glass tanks of 60 L. The system had water recirculation with controlled temperature, using an electronic heater at 27ºC.
For the adaptation of the juveniles to commercial food, the animals were subjected to feed acclimation when they were still in Furnas, and the feed was made of zooplankton.This procedure took 24 days and followed the methodology adapted to that used by Salaro et al. (2011), by which the animals were fed with successive six diets for four days each.The diets were offered in the following order: 100% beef heart; 80% beef heart plus 20% commercial feed; 60% beef heart plus 40% commercial feed; 40% beef heart plus 60% commercial feed; 20% beef heart plus 80% commercial feed; and 100% commercial feed (45% crude protein).
Pesq. agropec.bras., Brasília, v.53, n.6, p.664-672, June 2018 DOI: 10.1590/S0100-204X2018000600002After feed acclimation, 176 H. intermedius juveniles were randomly selected.Fish had an initial mass of 5.23±1.61g and a standard length of 6.70±0.72 cm.The animals were divided into 11 fish per aquarium, in glass tanks of 60 L (60 x 32 x 31 cm, length x width x height).All fish showed similar length and mass at the start of the trial.
The water system used an air compressor to recirculate with constant aeration, porous air stones, and temperature mantained at 27°C.The water quality parameters were monitored daily for temperature, with a mercury thermometer, and pH with the Labcon Test pH Tropical (Alcon Pet, Camboriú, SC, Brazil).Oxygen, ammonia, and nitrite were measured weekly with the Labcon Test for Toxic Ammonia, dissolved O 2 and nitrite NO 2 -(Alcon Pet, Camboriú, SC, Brazil).The values for pH, ammonia, oxygen, and nitrite were 9.03±0.55,0.003±0.000ppm, 6.5±1.0 ppm and 0.063±0.130ppm, respectively.
A completely randomized design with four treatments and four replicates was used, and the tanks were considered the experimental units.The treatments consisted of the following photoperiods: 0L:24D, 6L:18D, 12L:12D and 18L:6D, in which L is light hours, and D is dark hours.Fish were maintained at their respective photoperiods for 10 days for acclimation to the new condition and, subsequently, the experiment started and had 90 days of duration.
Fluorescent lamps of 40 W, with a light intensity of approximately 50.7 lx, measured with a digital luximeter MLM-1010 (Minipa, São Paulo, SP, Brazil) at the surface of the aquaria, were used for lighting in each treatment.The lamps were controlled by timers with the light period beginning at 7:00 h, except for the lighting program in which fish were kept in 0L:24D scheme, and the lamp remained off for all the experimental period.For illumination of the experimental room, a red lamp of 40 W was used to do all the handling of animals, without any interference in the photoperiod.Treatments were isolated from each other by masonry walls, and at the front by a blackout system, preventing light interference between them.
The animals were fed twice a day, at 8:00 and 17:00 h, with the commercial extruded feed Fri-Aqua (Trouw Nutrition Brasil, Campinas, SP, Brazil) containing 45% crude protein in pellets of approximately 2 mm diameter, at a ratio of 3% of body mass per day.The remained pellets were taken 30 min after feeding the animals.Pellets were then frozen and dried at 65°C for further calculation of feed consumption and feed conversion.
The aquaria were monitored daily, in order to verify the presence of dead animals.Fish that were not found in the aquaria, and fish found dead with evidence of cannibalism (absence of a body part), were considered dead by cannibalism.Dead fish without signs of cannibalism were considered only for survival rate calculation.For the recalculation of the amount of daily feed (3% of body mass), biometrics of all animals were performed at intervals of 15 days; the standard length was measured by an ictiometer, and the mass, with the aid of an analytical balance of 0.1 mg accuracy.In order to avoid the stress of this procedure, the animals were anesthetized by immersion in water with benzocaine (190 mg L -1 ) for approximately 2 min.
At the end of 90 days, all fish were fasted for 24 hours, anesthetized by immersion in water with benzocaine (190 mg L -1 ) for about 2 min, weighed, and measured (standard length).Then, 23 fish from each replicate were euthanized by immersion in water with benzocaine (250 mg L -1 ); skeletal muscle tissue samples were removed, and the carcasses, without the viscera, were placed in identified plastic bags which were taken to the freezer and, subsequently, sent for the procedures of body composition analyses.
In addition to these variables, the uniformity of mass and length of the tanks were calculated.To determine the uniformity of the aquarium, the following equation, proposed by Furuya et al. (1998), was used: , in which U is the uniform experimental unit, which represents the variation of mass or length (%); N t is the total number of fish in each experimental unit; and N ±20% is the number of animals that had 20% to more or less of the average mass and length for each experimental unit.This variables assess the uniformity on a scale of 0 to 100%; values closer to 100% represent a greater degree of uniformity.
In the histology of muscle tissue, the individuals selected for muscle fiber analysis were randomly selected.Approximately 2 cm of muscle tissue samples were collected from the epiaxial portion, near to the dorsal fin, totalizing 11 animals per treatment.The samples were immersed in isopentane, frozen in liquid nitrogen (-196°C), and stored in a freezer at -80°C.Later, cross sections of 10 μm thickness in a cryostat at -20°C were made and stained with hematoxylin-eosin.
Morphometric analyses of muscle fibers were performed.The smallest diameter of 200 white fibers of each animal was measured using a microscope connected to the computerized image analysis system AxioVision (Carl Zeiss Microscopy, Thornwood, NY, USA).Each muscle fiber was grouped into one of the three classes of diameter: <20 μm, 20-50 μm, and >50 μm (Almeida et al., 2008).The frequency of the muscle fibers was expressed as the number of fibers of each size class related to the total number of measured fibers.
In addition, the analysis of the body composition was performed.A total of 92 animals were used, and they were grouped according to the treatment and replicate to a specific pool.Each replicate was divided into two samples, which resulted in eight samples per treatment for the analyses of the body composition of the carcass: moisture; lipid (or ether extract); ash and crude protein, according to the AOAC International (Helrich, 1990).
The moisture content of fish carcasses was determined in kiln at 105°C for 24 hours until the obtention of constant mass.Lipids were determined by the Soxhlet method, by which 2 g of a sample were submerged in dry ethyl ether, leaving at reflux for 3 hours (Helrich, 1990).The samples were suspended to drain the excess solvent or oil for 30 min and, subsequently, with the recovery of the ether, only the lipids were left.The analysis of ash was done by the gravimetric method, based on the lost mass of the material subjected to heating at 550°C for 12 hours.For the crude protein content, measurements were performed by the micro Kjeldahl method (Helrich, 1990).
All animal experimental procedures in this study were approved by the animal ethics committee of Universidade Federal de Lavras (Standing Committees/ PRP-UFLA), and registered in the certificate number 004/14.
The variables were checked for normality by the Shapiro-Wilk's test and homoscedasticity of data variance by Levene's test.For the data considered as normally distributed, the one-way analysis of variance was applied, followed by the Tukey's test in case of significance, and for the data found to be not normally distributed, the Kruskal-Wallis' test was used.Additionally, a chi-square test was performed to check if the diameter of the fibers is dependent on the photoperiod.All analyses were performed with the program SPSS Statistics version 20, at 5% probability.

Results and Discussion
Photoperiod affected survival and cannibalism (p<0.05) of H. intermedius (Table 1).The 12L:12D photoperiod had the highest survival rate and no deaths by cannibalism.Additionally, the 12L:12D group showed high uniformity for length and mass of the animals.The opposite was observed for the 6L:18D treatment that showed the lowest survival rate and the highest cannibalism rate.The lowest uniformity was observed for both length and mass of H. intermedius in the 6L:18D group.Although the 0L:24D group showed higher uniformity values than the 12L:12D group, the first group had higher standard deviation than the second one, indicating a large variance.
The large size of heterogeneity is one of the causes of cannibalism, called cannibalism type II, in which the prey is completely consumed, generating significant high losses (Król & Zieliński, 2015;Naumowicz et al., 2017).This was observed in the 6L:18D treatment that had the highest cannibalism rate.The long dark period (18 hours), followed by a short light period (6 hours), probably stimulated the dominance behavior of some animals, once the species becomes more active at night and prefer areas of low light conditions (Bussing, 2002).These dominant animals could feed more, increasing their size in relation to the other animals.The appearance of these size differences between animals resulted in cannibalism, which led to a decreased density of the aquaria, contributing to the growth of those dominant animals.A consequence of Pesq. agropec. bras.,Brasília,v.53,n.6,p.664-672, June 2018 DOI: 10.1590/S0100-204X2018000600002 all these factors was an increase of the heterogeneity of the animals, which resulted in high cannibalism rates.
In the 12L:12D photoperiod, the great uniformity in the length and mass values of the animals causes the cannibalism decrease in this specie, thus favoring the survival (Luz et al., 2000).According to Reynalte-Tataje et al. (2002), there is a positive correlation between survival and the increase of the light period.In their study, Brycon orbignyanus larvae, a carnivorous fish at this stage of life, had the highest value of survival and uniformity among fish maintained under increasing periods of light (24 hours), and the lowest value of survival and heterogeneity of growth (because of cannibalism) when fish were subjected to a long dark period (0 hour).Our findings corroborate these data; however, for H. intermedius, it seems to exist a need for a long darkness period followed by short light stimulus to trigger cannibalism, since there was less survival and high cannibalism in the 6L:18D, but not in the 0L:24D treatment.Moreover, Salaro et al. (2011) found that H. intermedius juveniles, under different light intensities, including complete absence of light, have not been affected for their growth, survival, mortality, or cannibalism rates, during feed acclimation.Therefore, survival and cannibalism are possibly not affected by light intensity, but by photoperiod.However, Salaro et al. (2011) used a period of only 21 days in the experiment, which differ from this work, in which 90 days were used, representing another reason that may explain the differences in the survival and cannibalism rates.
Cannibalism and survival were also influenced by photoperiod in other species of carnivorous fish.S. scriptum larvae showed low survival to 0L:24D photoperiod due to high mortality and cannibalism (Schütz et al., 2008).According to these authors, the dark condition allows of the appearance of cannibalism and agonistic interactions.However, P. corruscans larvae showed high survival rate in the absence of light in the first five days, but later, under conditions of intermediate photoperiods (10L:14D, and 14L:10D), Table 1.Mean±standard deviation of survival, uniformity, growth, and body composition parameters of Hoplias intermedius juveniles maintained in photoperiod treatments during 90 days (1) .   Means followed by the equal letters in the rows do not differ at 5% probability by Tukey's test for survival and lipid on body composition; Kruskal-Wallis for cannibalism; one-way analysis of variance and Kruskal-Wallis' test for growth parameters. (3)The uniformity didn't have a p-value, since no statistical test was performed for this parameter.
fish showed the best survival rates (Campagnolo & Nuñer, 2008).Therefore, differences for survival of different species can be attributed to variations in the photoperiod requirement, which is species specific and depends on the development stage of the animal.The manipulation did not show influence on mass gain, length gain, specific growth rate, daily feed consumption, feed conversion and protein efficiency rate (p>0.05,Table 1).The same fact occurred in Lophiosilurus alexandri, that also showed no difference for growth when fish were maintained in different photoperiods (Kitagawa et al., 2015).The length and mass of B. orbignyanus postlarvae did not show significant differences between treatments (14L:10D, 10L:14D, 24L:0D, and 0L:24D), a great heterogeneity was observed between postlarvae cultivated in the condition of long darkness period (Reynalte-Tataje et al., 2002).
In the histology of muscle tissue, the average diameter of white muscle fibers showed a significant difference among photoperiod treatments (p<0.05).The animals from the 6L:18D and 18L:6D groups had the highest fiber diameter (56.36±25.85 and 56.08±28.91μm, mean±standard error, respectively), with no statistic difference between them.However, the 0L:24D and 12L:12D groups showed the lowest diameter values with 50.91±24.76and 50.65±24.33μm, mean±standard error, respectively, with no difference between them.
For each photoperiod, a low number of white muscle fibers (<20 μm diameter) and a large number of fibers with a diameter >50 μm were observed for all treatments, which indicated that the growth in this species occurs by hypertrophy (Figure 1).However, fish from the 6L:18D and 18L:6D photoperiods had greater degrees of distribution than those of the 0L:24D and 12L:12D (p<0.05)groups.The 6L:18D group showed 56.73% of the total fiber measures, and belongs to the class >50 μm, and the 18L:6D group showed 55.73% of the fibers in the same class.The 0L:24D and 12L:12D photoperiods had lower frequencies of fibers >50 μm, which were 50.13% and 47.32%, respectively.
The 6L:18D and 18L:6D groups were those with the largest diameters of muscle fibers and the highest frequencies in the number of fibers in class >50 μm, indicating that these photoperiods may have caused a great hypertrophy in the muscle tissue (Almeida et al., 2008).In other species, the continuous light resulted in high hypertrophic growth of muscle tissue in Hippoglossus hippoglossus, with the largest diameter of 50.71±1.24μm, mean±standard deviation (Lohne et al., 2012), and Umbrina cirrosa showed a great hypertrophy of white fibers in the 16L:8D photoperiod (Ayala et al., 2013).This may be due to the increased activity of animals during long photoperiods, which probably resulted in the high hypertrophy of white fibers, as occurred in the 18L:6D group in the present work.
A high cannibalism rate was observed in the 6L:18D group, which may have led to great hypertrophy of fish white fibers.This increase may be due to three factors: first, to the food type consumed, since digestible nutrients are often higher in fish that increase growth (Król & Zieliński, 2015); second, to the decreasing fish density in the aquaria, which also promotes the growth (Lima & Barbosa, 2016); and third, to the high fish activity due to chases and fights of cannibalism.This can also be observed by the low uniformity found in this group, both in fish mass and length.Therefore,

Figure 1 .
Figure 1.White muscle fiber diameter distribution in Hoplias intermedius juveniles subjected to photoperiod treatments during 90 days.Columns represent white fibers frequencies in each fiber diameter class.L:D, light:dark hours; *Significant difference at 5% probability from each other within the class >50 μm by the chi-square test.