Diet of fish ( Characiformes : Characidae ) during early developmental stages in the Upper Paraná river basin , Brazil

This study analyzed the diet during early developmental stages of Astyanax lacustris (AA), Piaractus mesopotamicus (PM), Megaleporinus obtusidens (MO) and Prochilodus lineatus (PL), under experimental conditions. Fish larvae, 350 of each species, were stocked separately in 16 fiber-cement tanks (500 L), from which, three larvae of each species were collected every three days, for 36 days. Tanks were fertilized (1.5 g NPK-7: 14: 8) and had 50% of the surface with macrophytes. Larvae were grouped in age classes = I: 5 to 11, II: 14 to 20, III: 23 to 29 and IV: 32 to 38 days and, dissected for analysis of the digestive tract. Methodologies of dominance, frequency of occurrence and points of food items were used. For AA and PM larvae, measurements were taken for the number and size of organisms to determine the percent participation in biovolume. Changes in diets of larvae in different age classes were found as well as between different species, with the same age. Rotifers were dominant in the content of digestive tracts of all species at Class I, and the larvae diets became more distinct with increasing age, in which AA consumed mainly rotifers, PM, larger organisms (mainly cladocerans), PL, algae (diatoms as the main item) and PP with a more diversified diet (rotifers, ostracodes and algae). In conclusion, early stages of these species presented distinct diets, undergoing remarkable changes in the first 38 days of life.


Introduction
Several fish species produce a large number of larvae in the breeding season, but survival is usually very low during this critical period.As a consequence, small changes in mortality rates can significantly affect recruitment of populations and this effect will be reflected in the subsequent years (Gerking, 1994, Osse, van den Boogaard, van Snik, & van der Sluys, 1997;Sánchez-Velasco, 1998;Rabe & Brown, 2000;Bunnell, González & Stein, 2003).For many fish species, the breeding season and site occur simultaneously with high densities of zooplankton, which is the main food resource for their young forms (James, Pitchford, & Brindle., 2003;Bremigan & Stein, 1999;Sammons, Bettoli, & Greear, 2001).The survival of these young forms depends on environmental conditions, availability of adequate food, and other biotic interactions (Miller, Crowder, Rice, & Marschall, 1988;Goshorn & Epifanio, 1991;Meng & Orsi, 1991;Rabe & Brown, 2000).Rapid growth of larvae is an important parameter to reduce loss by predation.For this, it is important to have an adequate diet that contributes with the necessary energy (Hartig & Werner, 1987;Miller, Crowder, & Binkowski, 1990;Pedersen, 1997).
In terms of feeding, fish larvae are essentially a 'species apart' when compared to their adults.In fish, larvae and adult are often completely different from each other, and may even be considered ecospecies, presenting peculiarities as to the type of habitat, feeding and behavior (Nakatani, Baumgartner, & Cavichioli, 1997, Osse et al., 1997).Newly hatched larvae are so small and undeveloped that their feeding ecology is greatly differentiated from the feeding ecology of the same species for the rest of the life cycle (Gerking, 1994;Osse et al., 1997).The micro-habitat used by the early stages of fish and their response to environmental variations is generally different from juveniles and adults of the same species (Childs, Clarkson, & Robinson, 1998).
Although the larvae of most fish species are very small compared to their adults, the larvae of different species may present different ecological responses to changes in abiotic and biotic variables.Some, because of size at hatch, swimming ability and capacity to utilize available resources, are more vulnerable to hunger, while others are more affected by predation and other biological factors (Miller et al., 1988).
Changes in the preferential food items of young fish stages depending on the species and also between individuals of different sizes in the initial phase, as well as diet plasticity of larger individuals, are reported for Colossoma macropomum and hybrid of C. macropomum ♀ x Piaractus mesopotamicus ♂ by Sipaúba-Tavares (1993), for the Rhinelepis aspera by Soares, Hayashi, Furuya, and Maranhão (1997) and for species of the genus Bryconamericus by Borges, Assakawa, Cunha, Bialetzki, andNakatani (2006), Suiberto, Galuch, Bialetzki, andNakatani (2009).
Unlike marine fish species, freshwater fish larvae depend more on small cladocerans than copepods, and rotifers are also common as the first food (Gerking, 1994;Soares et al., 1997).Larvae of tropical freshwater fish generally use rotifers as the first food, consuming cladocerans, nauplii, copepods and insect larvae, and present greater dietary plasticity as they develop (Miller et al., 1990;Soares et al., 1997, Suiberto et al., 2009).
This study aimed to characterize and compare the natural diet of the Characidae fishes Astyanax lacustris, Piaractus mesopotamicus, Megaleporinus obtusidens and Prochilodus lineatus in different stages of early development under experimental conditions.

Material and methods
Larvae of A. lacustris, P. mesopotamicus, M. obtusidens and P. lineatus witch 48 to 72 hours after hatching, were stocked at a density of 350 individuals per fiber-cement tank with 500 L capacity; 16 tanks were used, four for each species, for 36 days, during January and February 2001.Larvae were obtained by induced breeding in commercial fish farms with matrices and reproducers originating from Paraná River basin.
Tanks, coated with non-toxic water-based paint, were previously washed and dried, exposed to the sun, for two days and received water from artesian wells and treated, using 15 mL 12.6% sodium thiosulphate to inactivate chlorine.Fertilization with 1.5 g of inorganic fertilizer (NPK-7: 14: 8) was performed with weekly applications of 0.75 g NPK to each of the tanks.Tanks were inoculated with water from lakes of the Paraná River basin, in the region of Porto Rico, State of Paraná, seven days before larval storage to allow the proliferation of planktonic organisms.
Tanks were covered with 50% shading screen to prevent predator action.Aquatic macrophytes were placed, water hyacinth (Eichhornia crassipes) and water lettuce (Pistia stratiotes) in a proportion of 1: 1 in the tanks covering half of the surface, using wires fixed to the edges to delimit this space.
Every three days, three larvae were collected from each tank.These were fixed in 4% neutral formalin solution.The collected larvae were grouped into four classes, according to age as follows: I: 5 to 11 days, II: 14 to 20 days, III: 23 to 29 days and IV: 32 to 38 days.Subsequently, larvae were measured for total length under a stereomicroscope with a millimeter eyepiece.
For the analysis of the consumed organisms, the digestive tracts of the young stages were removed under a stereoscopic microscope, at magnifications of 10 and 40 X, with the aid of stilettos.Subsequently, with the appropriately separated content, under optical microscopy, the material was analyzed for number, frequency of occurrence, points and dominance, according to Hynes (1950).Analyses of the total content of the digestive tract were performed, except for larvae at the post-flexion stage, from which only the contents of the half or the initial 2/3 of the digestive tract were analyzed because of the lower degree of digestion of the organisms in these portions.For de A. lacustris and P. mesopotamicus larvae, we counted the organisms present to evaluate the percent participation of food items.
In the point methodology, a scoring scale was established according to the visual volumetric participation of each food item.In this way, the following procedure was adopted: points 1, 2, 3 and 4 for organisms that had a reduced participation (less than 5%), low (5 to 10%), moderate (10 to 30%) and important (above 30%) participation in the total volume analyzed for each fish, respectively.As in the case of the young stages of A. lacustris and P. mesopotamicus, a total count of the organisms present in the analyzed portions can be performed, measurements were taken total length, height and width (excluding spines) of at least 30 individuals of each food-organism for the determination of participation in biovolume of these in the digestive tubes of these fish.For the calculation of biovolume of food items, we used the methodology proposed by Ruttner-Kolisko (1977).

Results
There were differences in diets of larvae of different species with the same age and changes in diets of larvae of the same species in different age classes (Figure 1).In the first age class (5 to 11 days), there was predominance of rotifers in the diets of all species, with dominance in most of the gastrointestinal tract of these species.There was a predominance of Brachionus sp. both in dominance (Figure 1) and percentage of estimated biomass (Figure 2), in the contents of larvae of A. lacustris and P. mesopotamicus at this stage.However, for larvae of M. obtusidens and P. lineatus, there was a greater dominance of smaller organisms, such as small rotifers (Lecane bulla and Bdelloidea) and algae like Scenedesmus sp.(Figure 1).Regarding the score of food items (Figure 3), Brachionus sp. had the highest frequency (score 4) for larvae of A. lacustris and P. mesopotamicus, followed by L. bulla and cladoceran Macrothrix sp., respectively.For M. obtusidens and P. lineatus, higher frequencies of higher scores were found for Bdelloidea and L. bulla, respectively.It is noteworthy that there have already been differences in the diet of these fish in their first days of life.In the second age class (14 to 20 days), a more significant participation of cladocerans (Macrothrix sp. and Simocephalus sp.) was verified for all larvae, and Macrothrix exhibited the highest percentage of dominance in the content of the digestive tract (Figure 1).Cladocerans received the highest frequency of highest scores (score 4) for A. lacustris, P. mesopotamicus and P. lineatus (Figure 3), whereas for M. obtusidens, L. bulla was the item with the highest frequency of high scores.In the contents of A. lacustris, there was a greater participation of Chironomidae larvae and ostracods in terms of both dominance and biomass (Figure 2).The reduced dominance of rotifers and higher frequencies of lower scores for these were more evidenced for P. mesopotamicus followed by A. lacustris.
In the third age class (27 to 30 days), there were higher percentages of dominance and biomass in the contents (Figures 1 and 2) and higher scores in a greater percentage of gastrointestinal contents (Figure 3) of rotifers (Brachionus and L. bulla) and cladocerans (Macrothrix sp. and Simocephalus sp.) for A. lacustris and P. mesopotamicus, respectively.For M. obtusidens, the highest percentages of dominance and score 4 were observed for Brachionus sp., ostracods and periphyton protozoa, while for P. lineatus, the predominance was observed, based on the dominance and high scores of diatoms (unicellular and filamentous) typical of the periphyton.
The young stages with 32 to 38 days of age (fourth phase) presented distinct differences in the diets with predominance of rotifers (mainly L. bulla), cladocerans and diatoms in terms of percentages obtained for dominance and high scores for the young forms of A. lacustris, P. mesopotamicus and P. lineatus, respectively.On the other hand, M. obtusidens presented a differentiated diet with greater dominance of rotifers, especially Brachionus sp., however, with relevant participation for ostracods, unicellular and filamentous diatoms.These differences in the food items found in the diets of young forms were also evident with respect to the score obtained by these items in the gastrointestinal contents (Figure 3).
In relation to the frequency of occurrence of food items (Figure 4) in the gastrointestinal contents, changes in fish diets are also found both for the age of the same species and for groups of different species with the same age.

Food items Food items
Food items Food items In the last age class (32 to 38 days), we observed a clear difference between the items with the highest percentages of occurrence in the gastrointestinal contents of P. mesopotamicus and P. lineatus, with the predominance of cladocerans and algae (Coelastrum, Scenedesmus, unicellular and filamentous diatoms), respectively.In some cases, especially for A. lacustris, in the last age group, although some items presented high frequency of occurrence, they do not stand out in terms of dominance and scores in the contents and this is because, although they were present in many of the contents of digestive tracts, they were recorded in small quantities.

Discussion
The finding of the greatest importance of rotifers and cladocerans in the first 20 days of the larvae of the four fish species of this study is in accordance with the report of Gerking (1994), who considers that zooplankton is the main food for fish larvae and that this initial 'carnivorous' habit also occurs with those species that, when adult, are herbivorous or omnivorous, as is the case of the target species of this study.
To M. obtusidens in this work, L. bulla was the most dominant item.Sipaúba-Tavares (1993), with tambaqui larvae (Colossoma macropomum) and tambacu (C.macropomum ♀ x Piaractus mesopotamicus ♂), observed a greater dominance of rotifers in the first six days of life.Bremigan and Stein (1999) emphasize the importance of small zooplankton for the recruitment of Dorossoma cepedianum, observing a higher number of larvae in the weekly cohorts of larvae during periods with high biomass of small zooplankton, which is evidenced for fish in this study, especially in the first days of life.
In the first class (5 to 11 days), there were higher values of frequency of occurrence of rotifers mainly for A. lacustris and P. mesopotamicus, although, for P. lineatus, algae had already presented high values of occurrence.In the second and third classes (14 to 20 and 23 to 29 days of life), there were changes with increase in the frequency of occurrence of algae, mainly with the young stages of A. lacustris, M. obtusidens and P. lineatus; and for P. mesopotamicus, there were higher values for cladocerans.Furuya, Hayashi, Furuya, and Sakaguti (2002), on Pseudoplatystoma corruscans, Borges et al. (2006) and Suiberto et al. (2009), on species of the genus Bryconamericus, observed an increase in the frequency of occurrence of larger items (copepods, ostracods and insect larvae) and reduction in the occurrence of smaller items (rotifers), according to increasing age of the post-larvae, which was also observed by Galdioli, Hayashi, Soares, and Marques (2017) for Astyanax lacustris at early developmental stages.
Changes in the most important items were observed in the diets with increasing age of the young stages of all the species of this study, and the differences regarding the preferential items for each species became more evident in older individuals.Similar results were obtained by Sipaúba-Tavares (1993), with larvae of C. macropomum and hybrids of C. macropomum x P. mesopotamicus, which prefer smaller organisms, mainly rotifers at the beginning of development, starting to consume larger organisms (cladocerans and copepods) as they grow.In studies with A. lacustris, Galdioli et al. (2017), there were changes in the diet of these fish as they develop, with preference for Lecane sp.during the first days of development.
The importance of algae and rotifers in the diet of young stages of periphyton-feeding fish was reported by Soares et al. (1997), with Rhinelepis aspera, during the first 32 days of life, where rotifers and Scenedesmus spp were the most important items in the diet of this species.Shifts from more frequent items, as the fish grow, as obtained for the young stages of A. lacustris and P. mesopotamicus in this study, with reduction in Brachionus and Lecane percentage, to the increase in consumption of larger organisms such as cladocerans, were registered for larvae of I. labrosus and H. edentatus by Makrakis et al. (2005).
Differences in the diets of these fries, whether intraspecific or interspecific, should be related to changes in the morphology of both the digestive tract, ie due to differences in the size of the mouth, in the form and function of the stomach and in the length of the intestine, which influence the capacity of capture, digestion and use of the different diets, such as those that confer swimming ability and visual acuity.These factors can lead to species segregation by food and micro habitats.

Conclusion
The early stages of A. lacustris, P. mesopotamicus, M. obtusidens and P. lineatus present distinct diets, undergoing clear changes in the first 38 days of life.Rotifers represent the main item for the young stages of all species studied, with predominance of Brachionus for A. lacustris and P. mesopotamicus.As larvae of P. mesopotamicus grow, they begin to consume larger organisms, especially cladocerans, while larvae of A. lacustris and M. obtusidens consume rotifers and ostracods and P. lineatus consumes principally periphyton diatoms.

Figure 3 .
Figure 3. Percentages of scores of food items in the gastrointestinal contents of young stages of Astyanax lacustris, Piaractus mesopotamicus, Megaleporinus obtusidens and Prochilodus lineatus with different ages.
Percentages of dominant items in the gastrointestinal contents of young stages of Astyanax lacustris, Piaractus mesopotamicus Megaleporinus obtusidens and Prochilodus lineatus with different ages.