DIETS FOR FEED TRAINING PEACOCK BASS Cichla sp

The use of prized, carnivorous fish species such as the peacock bass Cichla sp. in either intensive farming or sport fishing demand specific knowledge on feed conditioning strategies for those species. One thousand and fifteen 0.5-g fingerlings were trained for 7 days to feed on ground fish flesh (GF). Seven hundred and seventy six (76%) fish (0.63 ± 0.03 g) feeding on GF were stocked into twelve 0.03-m net cages (63 fish/ cage) and submitted to gradual feed ingredient transition (GFIT) weaning technique. Moist pellets with 90, 80, 70 or 60% GF (GF-90, GF-80, GF-70 or GF-60, respectively) were offered during the first 4 days of GFIT. No fish accepted GF-00 at the end of GFIT. Fish started on GF-90 or GF-80 fed well until GF dietary levels dropped below 40%. To improve acceptance of pellets containing 30% or less GF, a second trial with four 0.03-m net cages stocked with 120, 0.5-g fish feeding on GF was designed. Fish fed on a sequence of moist pellets containing 90, 80, 70, 60, 50 or 40% GF for 3 days. Approximately 81% of the fish accepted GF-40; they were pooled and restocked into nine 0.03-m net cages and weaned to GF-00 with a sequence of diets containing 30, 20 and 10% GF plus: 1) a meat-flavored dry, commercial feed (MEAT); 2) a fish-flavored dry, commercial feed (FISH); or 3) MEAT plus 10% krill meal (KM). Fish accepted fish-flavored pellets better than meat-flavored pellets. Addition of 10% krill meal to a meat-flavored feed improved pellet acceptance, even when ground fish flesh comprised only 10% of the feed. However, no fish accepted GF-00 pellets at the end of this study.


INTRODUCTION
Brazilian aquaculture has experienced a significant growth in the last decade.Tropical climate in central, north and northeast Brazil, availability of land and water resources, production of grains and cereals for fish feeds, an enormous variety of fish species and a large, untapped domestic market for fishery products forecasts Brazil as a major supplier of cultured fish.Many Brazilian fish species with established markets are carnivores, and their intensive culture depends on the development of strategies to condition those species to accept commercial feeds.In addition, weaning juvenile, carnivore fish from natural food to dry pellets will provide research op-Scientia Agricola, v.60, n.4, p.609-613, Oct./Dec.2003 portunities to determine nutritional requirements for many species.
South American peacock bass Cichla sp., a Cichlid native to the Amazon River basin, fills an ecological position equivalent to largemouth bass Micropterus salmoides in North America.Both species have similar, piscivorous feeding preferences and general appearance.Feed training strategies for largemouth bass were developed in the United States (Snow, 1960;1963;1968;Kubitza & Lovshin 1997a;1997b).Strategies used to feed train largemouth bass may help conditioning Brazilian carnivores, including peacock bass, to accept dry diets.
Peacock bass can be potentially used either as sport, ornamental or food fish.The species has a white, firm-textured, boneless flesh and is considered one of the top table fish among Amazonian species.Peacock basses spawn naturally, several times a year in static water.While production of fingerlings in ponds is easily accomplished, few studies on weaning fingerling peacock bass from live food to formulated, dry feeds have been published (Fischer, 1991;Moura et al., 2000), and results were not very encouraging, ranging from 8 to 38% training success.The present study aimed to develop diets and feeding strategies for training peacock bass fingerlings to accept dry pellets.

MATERIAL AND METHODS
Peacock bass fry captured from a reservoir in the state of Goiás in Central Brazil were raised in fertilized ponds.One thousand and fifteen 0.5-g fingerlings were seined from the ponds and stocked into a 0.8-m 3 net-cage and fed with ground fish flesh to satiation, four to five times a day for 7 days.Seventy-six percent of these fish were trained to accept ground fish flesh, and were submitted to a series of trials aiming to wean them from ground fish flesh to a dry, pelleted diet.
Weaning diets were prepared by grinding and mixing a 30% crude protein, commercial, dry, fish-flavored feed with ground whiting flesh.All diets were supplemented with vitamin mix at 1% of total ingredients and were extruded on electric meat mincer.Diets containing over 60% ground fish had paste-like texture and were fed to fish after passing through a 1.5-mm screen sieve.Diets with 50% or less ground fish were presented to fish as 2 mm pellets.Dietary dry matter was determined after oven drying at 105 o C for 24 hours.Percent composition of the weaning diets is detailed in Table 2.
Fish were fed the experimental diets to satiation four times a day.Each weaning diet was administered to fish for four days.Percent feeders were determined at each diet change during the weaning period by dividing the number of feeders by the total number of fish stocked.Feeders were identified at the end of the day based on their larger, distended abdomen, contrasting with the smaller size and emaciated appearance of the non-feeders.Percent feeders were used as indicator of survival rates.Non-feeders were removed from experimental units at each diet change.

Trial 2
In Trial 1, fingerling peacock bass did not feed on pellets which have 30% or less ground fish flesh.Therefore, Trial 2 was set up aiming to enhance palatability of weaning diets with 30% or less ground fish and improve weaning success to a dry pellet without ground fish (GF-00).Peacock bass fingerlings which did not accept weaning diets in Trial 1 were pooled and maintained on a ground fish flesh diet for 5 additional days.At the end of this period, fish feeding on ground fish flesh (0.51 ± 0.02 g body weight) were stocked into four, 0.03-m3 net cages, 120 fish per cage.All fish were fed for three days with either GF-90, GF-80, GF-70, GF-60, GF-50 or GF-40 diets, in a gradual feed ingredient transition strategy.Percent feeders were determined at each diet change as previously described.Diets with 70% or more ground fish had paste-like texture and were fed as described in Trial 1. Diets with 60% or less ground fish were presented to the fish as 2-mm pellets.
Peacock bass feeding on GF-40 were pooled and restocked into nine, 0.04-m 3 net cages, 42 fish per cage.Fish were weaned from GF-40 to GF-00 using three diet sequences (n = 3), in a gradual feed ingredient transition (Table 3).Diets in the weaning sequence had 30, 20 and 10% ground fish and were flavored with either krill meal; a commercial, fish flavored feed or a commercial, meat flavored feed.Diets containing meat flavored feed served as control, since previous diets with 90 to 40% ground fish were prepared with the same commercial meat flavored feed.Diets with 40 to 90% ground fish had the same percent composition as diets used in Trial 1.The composition of diets with 10 to 30% ground fish is detailed in Table 4.The final dry pellet (GF-00) had the same composition as GF-00 used in Trial 1.

Trial 1
After four days of weaning from 100% GF (Table 5), the percentage of peacock basses accepting the prepared diets improved from 4 to 98% as the inclusion of ground fish flesh in the diets increased from 60 to 90% (P < 0.01).However, acceptance of pellets during weaning dropped significantly as the amount of ground fish flesh in the diets fell below 50% in GF-90 and 40% in GF-80 diets sequences.None of the weaning sequences yielded fish accepting dry pellets without fish flesh at the end of Trial 1 (Table 5).

Trial 2
Percent fish accepting prepared feeds decreased from 93 to 81% as the inclusion of ground fish flesh in diets was reduced from 90 to 40% (P < 0.01) (Table 6).When the amount of ground fish flesh dropped to 10%, the addition of 10% krill meal to the pellets maintained pellet acceptance at 81%, while only 15% and 4% of peacock bass accepted the fish-or meat-flavored pellets, respectively (P < 0.05).Fish accepted fish-flavored pellets better than meat-flavored pellets without krill meal (P < 0.05).Although 81% of fish accepted GF-10 pellets flavored with 10% krill meal, removal of krill meal and fish flesh from the meat flavored GF-00 reduced feed acceptance to nearly zero.Acceptance of the final dry pellets with no krill or ground fish ranged from 0 to 2% and showed no difference among the tested weaning sequences (P > 0.05).

Ground fish flesh as a starter diet for peacock bass
Ground fish flesh combines adequate flavor and texture and proved to be an effective starter diet to feed training peacock bass, as 76% of the fish accepted ground fish flesh after seven days of training.Fischer (1991) weaned 100% of forty eight, 0.4-g peacock basss from brine shrimp nauplii to ground fish flesh in 4 days.Snow (1960;1963;1968) found that 39 to 97% of largemouth bass accepted ground fish flesh as a starter diet.Kubitza & Lovshin (1997b) also reported 61 to 86% acceptance of ground fish flesh by 0.9 to 1g largemouth bass.
Weaning peacock bass from fish flesh to a dry pellet Fischer (1991) used gradual feed transition as a strategy to wean peacock bass fingerlings from ground fish flesh to moist pellet.Gradual feed transition consists of gradual replacement of ground fish flesh in moist pellets, feeding a progressively smaller amount of fish flesh and a larger amount of moist pellets.Weaning success was 38% when a commercial, microencapsulated diet was added to moist pellets as attractant, compared to only 8% for a moist pellet without attractant.Kubitza & Lovshin (1997a) found gradual feed ingredient transition more effective than gradual feed transition to wean largemouth bass from freeze dried krill to dry pellets.Gradual feed ingredient transition successfully weaned 64% to 91% of largemouth basses from ground fish flesh to dry trout pellets (Kubitza & Lovshin, 1997b).Detailed procedures for gradual feed transition and gradual feed ingredient transition are presented by Kubitza & Lovshin (1997a).Table 5 -Percent peacock bass fingerlings feeding on prepared feeds during 24 day of weaning from ground fish flesh to a dry pellet without fish flesh (GF-00).The weaning sequences started using diets with 90, 80, 70 or 60% ground fish (GF-90, GF-80, GF-70 or GF-60, respectively), followed by diets with decreasing amounts of ground fish, as detailed.
Table 6 -Percent peacock bass fingelings accepting weaning diets with 30% or less ground fish (GF) flavored with either krill meal (KM), fish-flavored commercial dry feed or meat-flavored commercial dry feed.GF-00 was a dry pellet without KM or GF1 .
1 Means in the same column not sharing the same letters are significantly different (LSD; P < 0.05).
2 Each diet was fed for 3 days.In the present study, the texture and flavor of the weaning diets were gradually altered as ground fish flesh was progressively replaced by ground, commercial feed during gradual feed ingredient transition.Lowering the content of ground fish in the first weaning diet in Trial 1 was an attempt to speed up the weaning process.However, this procedure caused sudden changes in dietary flavor and/or texture, which may explain the decrease from 98% to 4% feed acceptance by peacock bass as the inclusion of ground fish in the first weaning diet was reduced from 90 to 60% (Table 5).Kubitza & Lovshin (1997a) also observed a decrease in pellet acceptance after a sudden drop in the concentration of krill meal in the first feed used to wean fingerling largemouth bass from freeze dried krill to a dry pellet without krill meal.
Gradual feed ingredient transition worked relatively well as a weaning strategy for peacock bass.Sixty-six percent of peacock bass feeding on ground fish flesh accepted GF-30 when started on GF-90.Seventy-five percent of peacock basss were weaned from ground fish flesh to GF-40 when weaning began with GF-80.Decreasing the dietary levels of ground fish flesh below 30% resulted in poor feed acceptance.Dry pellets without ground fish flesh were not accepted by peacock bass in this study.Lowering ground fish flesh levels in diets to 40% or less coincided with 20 to 40% increase in the amount of fish meal added to diets.Most fish meal available in Brazil have peroxide values higher than 250 meq kg -1 , indicating a high degree of rancidity.Additionally, raw material used to produce these meals include fish processing wastes, resulting in high levels of biogenic amines.Even though detailed analysis of the sardine meal used in Trial 1 was not performed, it is fair to infer that rancidity and/or biogenic amines possibly played a role in lowering feed palatability and acceptance by peacock bass.

Enhancing feed palatability
Peacock bass fingerlings were gradually weaned from ground fish to a semi-moist pellet containing 40% ground fish flesh with a success rate of 81%.As observed in Trial 1, a sudden drop in feed acceptance was experienced when dietary ground fish flesh dropped below 40%.Except for pellets flavored with krill meal, pellet acceptance was further reduced when dietary levels of ground fish flesh dropped to 20%.Peacock bass accepted fish-flavored pellets better than meat-flavored pellets (P < 0.05).By the time peacock bass learned to accept moist pellets with 40% ground fish flesh, pellet flavor seemed more important than texture in determining the success of weaning to dry pellets.It shall be noted that hard pellets containing 10% ground fish flesh and flavored with krill meal were equally accepted or better than softer pellets containing 20% or 30% ground fish flesh without krill meal.Krill meal seemed to enhance pellet palatability for peacock bass.Removal of krill meal and ground fish flesh from the final dry pellet lowered feed acceptance to almost zero.Kubitza & Lovshin (1997a) found freeze dried krill and krill meal to be effective as attractants for feed training largemouth bass.Whole and crumbled freeze dried krill were accepted as a starter diet by 79 to 95% of 1-g largemouth bass.Krill meal enhanced flavor of dry pellets, increasing their acceptance by fish.Feeding response of red sea bream Pagrus major, Japanese eel Anguilla japonica, and gray mullet Mugil cephalus was improved when diets containing krill meal were offered (Allahpichay & Shimizu, 1984).
This preliminary study showed the possibility of weaning peacock bass fingerlings from live food to formulated feeds.Detailed studies are needed on the palatability of weaning and final dry diets to eliminate ground fish flesh from diets used to grow peacock bass to larger sizes.

Table 2 -
Percent composition of the weaning diets used during gradual feed ingredient transition for peacock bass.