FEEDING OF INTRODUCED SPECIES OF Cichla ( PERCIFORMES , CICHLIDAE ) IN VOLTA GRANDE RESERVOIR , RIVER GRANDE ( MG / SP )

Both species studied, Cichla cf. ocellaris and Cichla monoculus, fed mainly on fish, the first on Cichla spp., Plagioscion squamosissimus, and Tilapia rendalli, and the second, Cichla monoculus, on Cichla spp. Both diets indicated a strong correlation among the food items. Remarkable ontogenetic change was noted in both species diet: the young fed on crustaceans and insects while the adults fed mainly on fishes. The species studied, which are highly adapted to a brief life span, cause serious damage to the fish communities by predation, competition, and cascade effects throughout the whole trophic chain.


INTRODUCTION
Dam construction determines the course of river sectioning, in which running waters start to accumulate, resulting in lentic water masses.Obstactles to normal river flow.Intercalation of lentic environments along fluvial paths causes a fundamental transformation in hydrographic basins by changing hydrography, water quality, sedimentation, and aquatic life (Paiva, 1983).
Gallery forest is destroyed by overflow and many resources used by the fish communities become scarce, with alterations in occupation and growth occurring in species of aquatic plants (Castro & Arcifa, 1987).
Dams obstruct upstream migration of lotic fishes in addition to degrading habitats (Lowe-McConnell, 1990).
The reservoir has a great abundance of the tucunares Cichla cf.ocellaris and Cichla monoculus (Bloch & Schneider, 1801), fishes originally from the Amazon basin and having varied coloration and characteristic caudal occelus.Mandible protrusion, typical of this genus, ensures greater success in capturing prey with comparatively little energy expenditure (Motta, 1984).
The practice of introducing new species into recently created sites is common for a variety of reasons: recreational fishing, economic value increase, control of previously introduced species that have become undesirable, sentimental reasons, or for no reason at all (McDowall, 1968).Direct interactions between native and exotic species result from competition, predation, and parasitism (Vitousek et al., 1987).Many declines in fish population stem from overfishing, environmental changes, and introduction of species (Lowe-McConnell, 1990).
A successful species introduction may trigger extinctions and initiate a cascade effect in the whole community (Pimm, 1987).The most primitive forms become extinct when they lack a refuge to survive in places where more evolved species are established (McDowall, 1968).
Once established, a species is very difficult to eradicate and future effects of its continuation are unpredictable (Lowe-McConnell, 1990).
The tucunaré is a typical piscivorous fish, with direct impact observable in secondary consumer populations.Indirectly, other trophic levels are affected: the fishes fed upon by the predator and, in turn, ichthyophagous birds, insects, and the zooplankton (Zaret & Paine, 1973).Regardless of these consequences, great support exists in some sectors of a given community for introducing the tucunaré because of its huge potential in recreational fishing (Machado-Allison, 1971;Novoa et al., 1990;Larrazábal, 1996;Shafland, 1996).
In Florida, the tucunaré (Cichla sp.) was introduced, but not until intensive research was done to verify the probable effects of colonization.With respect to species introductions in general, Courtenay & Robins (1973) pointed out that in unaltered environments, competition from native species may exclude exotic fishes, which depends on many factors including the number of specimens introduced, their adaptations, and adaptability.In some cases, extinction of native species may occur, but this biomass loss is compensated for by exotic species population growth.Shafland (1996) overviewed the tucunaré introduction in Florida and found that the species adapted well, becoming very abundant and acquiring great socioeconomic value in fishing.However, the introduction of this species must be regarded cautiously, and caution should be exercised in new introductions.
The objective of this work is to characterize the feeding of both tucunare species (Cichla cf.ocellaris and Cichla monoculus) introduced in Volta Grande reservoir.

MATERIAL AND METHODS
Collections were made in April and June 1997 and monthly from September 1997 to August 1998, using two kinds of fishing equipment: gill nets with meshes of 2, 3, 4, 5, and 6 cm (between adjacent knots); and fishing rods with natural and artificial bait.Each gill net set was 150 m in length and 1.5 m in height, totaling 750 m.
For every specimen of each species, data were taken: total weight (TW) in grams, total length (TL) in centimeters, repletion degree of the stomach (RD), fat degree in the visceral cavity, and sex.
The stomachs with category 3 RD were withdrawn from the visceral cavity, weighted, and fixed in formalin 5%.After some days in formalin, the stomach content was transferred to alcohol diluted at 70% (Zavala-Camin, 1996).For stomach content analysis, 76 stomachs of Cichla cf.ocellaris and 18 of Cichla monoculus were used.Alimentary item identification occurred to the lowest taxonomic level reached.In addition, whenever possible the following were determined: number, weight, and length of each alimentary item.Preys found were analyzed using the following indexes (Matallanas, 1980(Matallanas, , 1982a, b;, b;Braga & Braga, 1987): • Numeric Index (Ni): relation in percentage between the number of preys from a taxonomic group and the total number of preys; • Frequency Index (Fi): relation in percentage between the number of stomachs with preys from a determined taxonomic group and the total number of stomachs with food; • Frequency in weight (Wi): relation between the weight of the preys from a taxonomic group and the tot al weight of all preys.
Through the (Ni) and (Wi) data the alimentary coefficient (Q) was obtained by the formula: Q = Ni .Wi.This coefficient is classified in three categories: Q more than 200 (preferential prey), Q between 20 and 200 (secondary prey), and less than 20 (occasional prey) (Matallanas, 1981).
To analyze the alimentary items found, the alimentary preference degree (APD), was used following Braga (1999) in accordance with the formula: APD = Si/N, where Si is the sum of the values related to abundance of the food item (i) in the stomachs and N, the total number of the stomachs analyzed.The estimated value for the alimentary preference degree (APD) for each item is indicated by: • APD = 4: the chosen item has absolute preference; • 3 < APD < 4: the chosen item has high preference degree; • 2 < APD < 3: the chosen item is preferential, otherwise different items are ingested, in which case; • 1 < APD < 2: the chosen item is secondary; • 0 < APD < 1: the chosen item is occasional.
Feeding variation related to the length was analyzed, and fishes were separated in length classes.The average number of preys (aN) and their average weight (aW) were calculated for each class.
Seasonal alimentary variation during the collection period was also analyzed.

RESULTS
The stomach repletion distribution demonstrated that in both species level 1 predominated, with the exception of the October collection that revealed level 2 for Cichla cf.ocellaris (Figs. 1 and 2).
The feeding coefficient analysis (Q) revealed that for Cichla cf.ocellaris the item Cichla sp. was preferential and the items Plagioscion squamosissimus and Tilapia rendalli were secondary, the others being occasional.Similarly, for Cichla monoculus the item Cichla sp. was preferential, and only Tilapia rendalli was secondary; the other items were occasional (Figs. 3 and 4).
The average number (aN) and average weight (aW) of the preys ingested by both species revealed that for Cichla cf.ocellaris, the average number is high in the small classes.On the other hand, the average weight of prey is the opposite, with higher values in the bigger length classes.In Cichla monoculus the number and weight averages increase in the higher classes (Figs. 5 and 6).
According to the alimentary preference degree (APD) of Cichla cf.ocellaris, the item Cichla sp. is secondary, reaching a value equal to 1.63, with all the others being occasional.For Cichla monoculus this item is preferential, having a value equal to 2, with all others being occasional.
Both diets were analyzed according to Spearman's rank correlation, which revealed that the diets are similar (p < 0.05).

DISCUSSION
The low quantity of replete stomachs sampled was expected, as they are of carnivorous species, as noted by Braga (1998).
The feeding of both species demonstrated that piscivorous behavior was the rule, although crustaceans and insects were also included; little diversity of alimentary items was found.A study carried out in Volta Grande reservoir also showed a low number of preyed-upon species (Braga & Gomiero, 1997).
The feeding spectra for the Cichla genus is based mainly on fishes, however such other items as aquatic insects and crustaceans may occur secondarily or frequently be ingested by young specimens.According to Zaret & Rand (1971)    The average number of preys was higher in the small length classes of Cichla cf.ocellaris due to a great Culicidae larvae occurrence, although this item contributed very little in weight.
Tucunarés capture their prey by chasing, and act like selective agents in removing young preys (Helfman, 1986).
According to Lowe-McConnell (1975), this genus is specialized in fish eating, with little dietary difference throughout the year.Arcifa & Meschiatti (1993) also concluded that Cichla ocellaris is piscivorous, mainly finding in their stomachs Cichla (48-57 mm total length) juveniles.In tucunarés of less than 86 mm TL, their findings principally showed aquatic insects and, in lesser proportions, zooplankton and fishes.
Many predator species are characterized by ontogenetic changes in diet related to changes in: morphologic characteristics, predatory behavior, habitat selection, and prey availability (Schmitt & Holbrook, 1984).During and after the parental care period, tucunaré juveniles (until 80 mm) feed themselves on zooplankton, rather than cladocerans, copepods, and microcrustacea.Above this length, the diet is based on crustaceans, insects, and small fishes, with fish the preferential item in adults (Lowe-McConnell, 1969, 1999;Zaret, 1980;Santos et al., 1994;Jepsen et al., 1997;Winemiller, 2001).Futuyma (1992) showed that the predators in varying degrees focus their food search efforts on the most common kinds of prey, even shifting to places where a particular kind of food is concentrated.By acting as if they were equipped with a search image, predators both increase the effectiveness of the search and decrease impact on uncommon kinds of prey.
The fishes consumed by tucunaré vary according to the environment, with the greatest predation incidence among the most abundant species.Fish of their own genus, forage fishes (Characiformes), and bentonics (pimelodids and ciclids) are frequent alimentary items (Novoa et al., 1990;Santos et al., 1994;Gil et al., 1996;Novoa, 1996;Jepsen et al., 1997;Williams et al., 1998;Lowe-McConnell, 1999).Ogari (1985) studied the prey-predator relation in Lake Victoria, with emphasis on Nile perch (Lates niloticus Linnaeus, 1758).After stabilization, this predatory species presented a high cannibalism degree and considerably reduced the endemic species stock.After introduction into the most varied environments of species from the genus Cichla, a drastic alteration was observed in the fish community.The small-fish population began decreasing due to tucunaré predation, which can ultimately bring about alterations at all trophic levels (Arcifa & Meschiatti, 1993;Santos et al., 1994;Santos & Formagio, 2000).When the tucunaré is introduced into favorable conditions, what may result is the extinction of some autochthonous species through predation (Zaret & Paine, 1973;Godinho & Formagio, 1992;Molina et al., 1996).On the other hand, this threat may be offset by dietary flexibility of endemic tropical piscivores (competitors of the introduced piscivores), which in these environments may be contributing to assure population maintenance (Pompeu & Godinho, 2001).
According to Godinho et al. (1994), the richness of species in lakes following tucunaré and piranha introduction declined almost 50% in relation to what it had been and was smaller than that found on lakes without introductions.The number of captured fishes was not significantly different between the two groups of lakes, but the biomass was higher in lakes with introductions.These lakes also differ from the lake group without introductions, in which small fishes were abundant.
Tucunarés do not cause overbalance in their natural environment owing to the adaptations of countless species of prey to predatory pressure and to the great number of competing piscivorous species.According to Lowe-McConnell (1999), great predator diversity maintains high prey-species diversity by allowing coexistence, thus keeping the predator number lower than it would otherwise be if these species were competing with one another for food or space.
The introduction of Cichla mainly occurs in altered environments, where the community of fishes is already in decline.The presence of these highly adapted and quickly proliferating predators causes serious damage to these communities by predation, competition, and cascade effects throughout the whole trophic chain.However, tucunarés attract recreational fishermen, involving in this way the whole riverine community that depends on native fish species.This factor explains recent introductions in these altered environments.

Fig. 5 -
Fig. 5 -Average number and weight of the preys found in the stomachs of Cichla cf.ocellaris analyzed by length classes.

Fig. 6 -
Fig. 6 -Average number and weight of the preys found in the stomachs of Cichla monoculus analyzed by length classes.

TABLE 1 Alimentary items used by Cichla cf. ocellaris and Cichla monoculus introduced in Volta Grande Reservoir.
, low food abundance means more competition among the fishes.