Diet composition of Plagioscion squamosissimus ( Heckel , 1840 ) , a fish introduced into the Tietê River system

Plagioscion squamosissimus is a species from the Amazon basin that was introduced into the Tietê River system. The present study aimed to analyse the feeding habits of this species in Bariri Reservoir and to verify the possible occurrence of ontogenetic changes in its diet composition. The samples were gathered in four periods of the year: February, June, September and November 2003. The fish were gathered with different fishing net meshes in three different reservoir portions. The alimentary items found in the stomachs were identified to the lowest possible taxonomic level and had abundance, occurrence frequency, volume and biomass determined. The Alimentary Index (IAi) was calculated for each alimentary item consumed by “corvina” for each studied period. Comparisons among the diet of different size classes of P. squamosissimus were done using the similarity coefficient of Jaccard and the Cluster Analysis (UPGMA). The Friedman proof was performed to verify if there is a significant ontogenetic variation in the species diet and changes in the consumption of different alimentary categories by P. squamosissimus among the sampled periods. P. squamosissimus presented a piscivorous feeding habit, although other items were also consumed. The biggest values of IAi were obtained for the alimentary item fish in the months of June (0.47) and November (0.39). The item Ephemeroptera (Campsurinae) was the most representative in February (0.30) and June (0.45). Despite the fact that P. squamosissimus consumed an ample spectrum of alimentary items, the ontogenetic changes were evident through the exploration of aquatic insects by the younger classes and by a diet mainly composed of fish in adult individuals. The alimentary plasticity of P. squamosissimus evidenced in this study might have contributed to the success of this species in Bariri Reservoir.

The aim of the present study was to characterise the food consumed by Plagioscion squamosissimus ("corvina") and to determine whether there are changes in the items consumed attributable to ontogenetic changes, in Bariri Reservoir, Middle Tietê River basin, São Paulo State, Brazil.

Study Area
Álvaro de Souza Lima (Bariri) reservoir was formed by the damming of the Tietê River in 1965.It is the second from top of a cascade of six reservoirs in the system (Figure 1).It is located at 22° 06' S and 48° 45' W, at 420 m altitude (Cesp, 1998).
The reservoir has: 5,546 ha flooded area; 35,430 km 2 catchment area; 8.6 m mean depth; 7-24 days residence time and an annual mean discharge of 443 m 3 /s.It belongs to the Tietê basin, a part of La Plata basin.The reservoir is located in the northwestern part of São Paulo State, between the towns Bariri and Boracéia.

Methods
Sampling was carried out in four periods of the year: February, June, September and November of 2003, which cover the dry (June/September) and wet (February/ November) seasons.Fishes were sampled in three portions of the reservoir: upper, middle and lower (near the dam).At each sampling site, three batteries of gill nets were used, containing mesh sizes of: 3, 4, 6, 8, 10 and 11 cm opposite knots.Nets were placed in the reservoir and left for 12 hours.Species caught and respective numerical abundances were recorded in the field and specimens were then preserved in 10% formaldehyde solution.At the laboratory they were measured (total and standard length) and weighed.
Fish stomachs were removed, weighed and their content preserved.Food items were observed under stereomicroscope and identified down to the lowest taxonomical level possible, using available keys and with the help of specialists.
The method of Hynes (1950) was used to express the frequency of occurrence of the different food items in P. squamosissimus diet.

Introduction
The presence of humans in natural environments leads, in practically all cases, to the introduction of animal and plant species, whether intentionally or accidentally.In either case, the introduction leads to partial or total extinction of native species or is immediately followed by alterations in the trophic chain, balance of natural populations, or ecological processes (Espindola et al., 2003;Rocha et al., 2005).
Among animals, fishes are the most frequently introduced species, being transplanted from one hydrographic basin to others, in the same or even different continents (Rocha et al., 2005).In the reservoirs of the Tietê River, São Paulo State, Brazil, a number of species have been introduced (Smith et al., 2005).One of these, Plagioscion squamosissimus (Heckel, 1840), native to the Amazon basin, was introduced in São Paulo in 1966 by the Hydroelectric Company, initially into the rivers Pardo, Grande and Paraná, as well as the Ilha Solteira and Jupiá reservoirs, from which they colonised the Tietê River reservoirs (Braga, 1998;Agostinho and Júlio Jr., 1999).
Transplanting species from one basin to the other can represent a menace to the local species.Introduced fish may be better adapted to or more tolerant of the environmental conditions, thus achieving higher growth and reproductive output (Espindola et al., 2003).Some introduced species may hybridise with native closely related species, competing for food and space for reproduction.They can also carry disease agents (pathogens and parasites) (Rocha et al., 2005).Competition for food is, however, the main way in which introduced species affect native ones (Agostinho et al., 1994).
From data on the feeding habits of fishes, including not only the composition but also the abundance of food items, it is possible to identify trophic categories and draw inferences about community structure that shed light on inter-relationships between components of the community (Agostinho et al., 1997).
Trophic ecology studies reveal a considerable plasticity in the diet of most teleosts (Abelha et al., 2001).Most fishes can switch from one food item to another as soon as alterations in food availability occur (Hahn et al., 1997;Agostinho and Júlio Jr., 1999;Wooton, 1999), so that diet composition may reflect food availability in the environment (Wooton, 1999).
Dietary changes may be related to seasonality in food availability, ontogenetic changes or changes in the food spectrum (niche width) in the presence of other species (Lowe-McConnell, 1987).
Friedman Test (Siegel and Castellan, 2006) was applied to the occurrence frequency of food items in order to verify if there were ontogenic changes in the species dietary.The null hypothesis tested is that there is no difference among the species population size classes regarding the food categories consumed.The alternative hypothesis is that all size classes present preference for some food categories.Only those size classes that consumed at least three food categories were included in the test.This same test was used in order to check if there were significant seasonal differences in the food categories consumed by P. squamosissimus regarding the different sampling periods in the Bariri reservoir.The null hypothesis is that there are no differences in the preference by food categories when comparing the items consumed in the different periods.The alternative hypothesis is that there are significant differences in the food categories consumed by P. squamosissimus population in the different periods of the year.The significance of the values obtained was verified using the chi-square test of critical values.

Results
Stomach contents of 258 specimens of Plagioscion squamosissimus were analysed and 184 stomachs were found to contain food, while 74 were empty.The overall pattern of frequency of occurrence varied markedly (Table 1).Ephemeroptera (Campsurinae fragments) were Food item volumes were obtained by volume displacement, using 10 and 30 mL graduated cylinders (Hyslop, 1980).The relative contributions of the individual food items were also determined as biomass dry weight, by the gravimetric method (Hyslop, 1980).The dry weight of Diptera consumed was estimated with the conversion factor 0.0063 mg (Chironomidae weight value), from Prat and Rieradevall (1995).
In order to evaluate the relative importance of each item in the diet of P. squamosissimus in the four studied periods, the percentage values obtained for the frequency of occurrence (%F) and biomass (%P) were combined in an Alimentary Index (IAi) as suggested by Kawakami and Vazzoler (1980), although replacing volume by the dry weight of each item.
In order to analyse if there were significant seasonal differences in the diet among P. squamosissimus size classes and among the four periods of the year in Bariri Reservoir, food items were grouped in the following categories: Fish, Ephemeroptera, Odonata, Trichoptera, Crustacea, Diptera, Insect fragments, Plant material and Detritus.
Ontogenetic differences in P. squamosissimus diet were analysed by separating the individual in different size classes (standard length) by Sturges method (1926).The seven size classes were compared with regard to the presence or absence of food items, by using the Jaccard index and then by the Cluster Analysis (UPGMA).The ume consumed, followed by insects (44.8 mL; 27.2%) (Table 1).
In terms of dry weight, the relative importance of each food item was fish (13,530 mg) representing 91.1% of total biomass followed by insects (5,120 mg; 34.4%) (Table 1).ingested by 36.4% of all specimens whereas fish (including both flesh fragments and whole fish consumed) were ingested by 47.6% of the P. squamosissimus analysed, these being the most frequent items consumed.
The most important food item in terms of volume was fish (116.8 mL), representing 71% of the total vol- The highest values of the Alimentary Index were found in June (0.47)and November (0.39) for the item fish flesh fragments.On the other hand the item Ephemeroptera (Campsurinae nymphs) was the most representative in the months February and June with 0.30 and 0.45, respectively (Table 2).
The population of P. squamosissimus was divided into 7 standard length classes (Class 1: 8.3 to 12.3 cm, Fish was the food item consumed in the greatest number.In all, 129 whole fish were found in the stomachs of P. squamosissimus analysed, followed by dipteran pupae and larvae, which totalled 71 organisms (Table 1).
Comparing the food items consumed in different periods of the year (Figure 2) it was found that in February the most frequent items were Campsurinae (Ephemeroptera nymphs) (62.2%), egg-mass Ephemeroptera (37.7%), fish flesh (26.2%), white fish (16.3%) and plant material (13.1%); whereas in June fish flesh became the most important item, with a frequency of occurrence of 33.8%, followed by Campsurinae (Ephemeroptera nymphs) (32.2%).In September there was a change in diet with predominance of Campsurinae (Ephemeroptera nymphs)  3. It was found that in February (end of summer and rainy season) fish, Ephemeroptera and plant material were consumed by all size classes, except class 5, which consumed plant material only.In June ( winter and dry season), individuals belonging to the classes 1, 3 and 4 consumed mainly fishes, Ephemeroptera, Diptera and insect fragments.Individuals from class 2 consumed only Ephemeroptera, Trichoptera and insect fragments.Larger individuals consumed only the item fish.In September (early spring and at the end of the dry season), individuals from class 1 and 2 consumed mainly Diptera, zooplankton (Copepoda) and plant material.
The item fish was consumed by classes 3, 4, 6 and 7.
It was observed that the fish belonging to the smaller size classes had no preference for any specific food, being characterised by a generalist feeding habit.On the other hand, the adult fish from larger classes the displayed a preference for the items fish flesh, indicating a more specialized diet.
Comparing the diet of the specimens of each size class sampled in each season (four sampling periods) is pre-  constituted a small fraction in the diet of the "corvina" (P.squamosissimus) in many water bodies in the Amazon region, being crustaceans the dominant item (Goulding and Ferreira, 1984).
The variety of items found in the diet of P. squamosissimus in Bariri Reservoir indicates a wide spectrum of food items consumed and consequently a low feeding specificity.According to Wootton (1999) and Abelha et al., (2001) classification of fishes feeding habits this species can be considered a generalist with a wide food spectrum.
Although most individuals consumed a variety of food resources, there was a noticeable shift in their diet during ontogenetic development.These changes consisted mainly of the greater exploitation of Diptera larvae (Chaoboridae and Chironomidae) and other aquatic insects (immature forms) by the juveniles of the smaller size classes and the switch to a fish dominated diet in the adults, characterising a wide spectrum of food items in the young and a more specialised, piscivorous diet in the adults.
The consumption of chironomids gradually decreased with the increasing size of P. squamosissimus specimens.As predacions fish grow, feeding on small organisms becomes less advantageous (Hynes, 1950).The capture of such organisms, unless very abundant, results in low forage efficiency and energy return.According to Wootton (1999), the ontogenetic changes in fish diet are also determined by morphological changes and greater locomotory ability in the larger fish.
Similar ontogenetic changes in the diet of P. squamosissimus were reported by Hahn et al. (1997), in a stretch of Rio Paraná (Porto Rico region, Paraná, Brazil) with juveniles mainly consuming insects and adults preferring to catch fish.
Besides fish and aquatic insects, plant material was an important resource used by P. squamosissimus, particularly in the months of February and November.The ingestion of this resource in these months can be influenced by the increase in the rainfall precipitation and greater input to the reservoir.Allochtonous plant matter (leaves, flowers, roots, and fruits) falling onto the surface can be carried to the reservoir during the rainy season.Uieda and Kikuchi (1995), studying the input of allochtonous material in a creek, concluded that the input occurs in the rainy season.
The food plasticity of P. squamosissimus evidenced in the present study might at least partially explain the reasons why this species has been so successful in this Reservoir.
Although it is generally believed that Plagioscion squamosissimus would be a good option regarding the selection of a species to be introduced in reservoirs, because it occupies mainly the deep limnetic region, usually not used by other fish, our study shows that actually, both juveniles and adults can compete for food, to the detriment of several native species populations.For this reason, P. squamosissimus should not be introduced into Individuals from class 4 were the only ones consuming the item Crustacea (Macrobrachium sp).In November Diptera and Ephemeroptera were consumed only by the smaller individuals (classes 1 and 2).The item fish was consumed by classes 1, 2 and 3, whereas class 4, only plant material (Figure 3).
Statistical analysis showed that there were significant differences in the food items consumed along the year in Bariri Reservoir.The null hypothesis was rejected (Xr 2 = 3.72; P < 0.05).
A dendrogram obtained by cluster analysis (UPGMA) of the presence or absence of the food items consumed by all 7 size classes of P. squamosissimus in Bariri Reservoir showed two distinct groups, one formed by the small size classes (1, 2, 3 and 4) and the second formed by the large ones (5, 6 and 7) (Figure 4).
The results from the Friedman Test applied to compare the diet of the different size classes also evidenced that there were significant changes in the food preference among the size classes and therefore the null hypothesis was rejected (Xr 2 = 6.48;P < 0.05).

Discussion
On the basis of the results obtained in Bariri reservoir, P. squamosissimus can be considered a fish of piscivorous feeding habit.Nevertheless, it also consumes a number of other food items, including many aquatic insects (Ephemeroptera, Odonata, Trichoptera e Diptera) and plant material.
This species was classified as piscivorous and generalist by Hahn et al. (1997), who also noted the occurrence of insects as secondary items.According to Braga (1998), the diet composition of P. squamosissimus in Barra Bonita Reservoir is essentialy fish, with a minor consumption of Odonata and Ephemeroptera nymphs.
In the Northeast of Brazil the food of P. squamosissimus is mainly based on fish, insects and crustaceans (Braga, 1990).On the other hand fishes basins other than those to which they are native, as for example that of the Paraná River.

Figure 1 .
Figure 1.Map of São Paulo State showing the location of Bariri Dam and a sketch showing the cascade of reservoirs on Tietê River, São Paulo state, Brazil.

Figure 3 .
Figure 3. Main food items consumed by different size classes in the Plagioscion squamosissimus population in Bariri Resevoir, Tietê River Basin, São Paulo State, in each of the four periods sampled in the year 2003.

Table 1 .
Contribution by abundance, frequency of occurrence (%FO), displaced volume (mL) and dry weight (g) of food items to the overall diet of Plagioscion squamosissimus in sampled in four periods of the year 2003 (February, June, September and November) in Bariri reservoir, Tietê River basin, São Paulo State, Brazil.

Table 2 .
Values of the Alimentary Index (IAi) for items of the P.squamosissimus diet in four periods of the year 2003 ( February, June, September and November) in Bariri reservoir, Tietê River basin, São Paulo State, Brazil.