EFFECT OF DIFFERENT PROTEIN LEVELS ON THE GROWTH AND FEED EFFICIENCY OF SILVER BARB ( Barbodes gonionotusl IN FERTILIZED PONDS

This study proposed to establish a suitable protein level in feed formulation for silver barb in a semi-intensive pond. Fingerlings of fish of approximate one-month age were raised in floating cages set in fertilized ponds at a density of 36 fish per cage. Three isocaloric diets with different protein levels (20, 25 and 30%) were given to the fish at ad libitum twice a day for 30 days of the experimental period. The results indicated that an optimal dietary protein level for growth and feed eficiency of silver barb was about25o/o This dietary protein level resulted in low FCR. Therefore, it can be concluded that the diet for fingerlings of silver barb in fertilized ponds should be formulated to contain 25ok protein.


INTRODUCTION
The herbivorous fish species, silver barb ortawes, Barbodes gonionotus, has been cultured semi-intensively in many Asian countries.Various feed ingredi- ents and dietary protein levels were encouraged in Thailand (Sattar, 1987).According to Somsueb (1993), feed for herbivorous species was formulated with 30% fish meal, 45o/ofine rice bran,24o/opeanut mealand 1% vitamin and mineral premix to contain 30% pro- tein.ln Cambodia, combination of organicfertilizerand supplementary feed of rice bran, corn meal, termites, red ants or aquatic vegetation was applied to silver barb ponds (Nuov & Nandeesha, 1993).In addition, commercial feed tends to be used in fish culture prac- tices.However, feed in any form contributes the high- est cost to total fish production.Reducing feed cost is encouraged.In this case, feed formulation for cultured fish in ponds needs to be determined (Wee,   1 e8e).
In feed formulation, dietary protein for maximum potential growth rate of cultured fish needs to be determined (De Silva, 1989) because this ingredient plays a vital role in fish growth but is expensive.A wide range of dietary protein levels for maximum growth rate of herbivorous fish under laboratory conditions has been observed.Wee and Ngamsnae (1987) recommended about 35% dietary protein level forsilver barb fingerlings.The dietary protein level for other herbivorous species such as Oreochromis aureus (Davis & Stickney, 1978); O. niloticus (De Silva &   Perera, 1985); and Z Zillii (Mazid et al.,1979) was about 30%.A study by De Silva & Gunasekera (1992)  showed that different major carps in spite of their varying food habits required dietary protein at a level of 30%.Inclusion of natural foods affects the composition and amount of diet required by fish (Hepher, 1990;   NRC, 1983) since they contain high protein, a mate- rial similarto growth promoters, vitamins, and miner- als (Viola, 1990).However, there is no information of nutrient requirement for silver barb in ponds.The ob- jective of this experiment, therefore, is to evaluate the effect of dietary protein on the growth and feed utilization of silver barb cultured in a fertilized pond.

Feed and Feeding
Three isocaloric feeds containing different protein levels of 20 , 25 and 30% respectively (Table 1) were tested in this experiment.The main source of protein in all feeds was soybean meal.Fish meal, rice bran, cassava starch.corn oil and fish oil were added in diet formulation.Commercial vitamin in this experimentwas bought from CP.This mineral premix was adopted from NRC (1983).
Feeds were prepared to cover experimental feed requirements for two weeks.Feed ingredients were analyzed fortheir proximate composition.In addition, the quality of these feedstuffs was observed by smell, texture and color.Feed ingredients were mixed and thereafter cooked cassava starch was added.The resulting dough was converted into pellets using a Hobart-food mincer.Feeds were then dried.Dried feed was re-analyzed for proximate composition and stored in plastic bags.
Dry matterwas determined by drying the sample in an oven at 105"C to reach a constant weight.Crude protein was determined using the Kjeldahl method in  (Iecator, 1978) by determining the total nitrogen content.The obtained value was multiplied by a factor 6.25 (AOAC, 1985).Crude lipid was determined using solvent extraction method in Soxtec system HT 1046 Service Unit (Tecator,1978).Crude fiberwas determined by acid- based digestion method using a Fibertec system (Tecator, 1 978).Ash content was determined by combustion of the sample in a muffle furnace at 550"C for 6 hours.Gross energy was estimated using the multification factors 5.65, 9.45, and 4.2 kcal/g for pro- tein, fat and carbohydrate, respectively.
Experimental feeds were given to the fish ad libiturn twice a day (about 09,00 and 17.00) for the 30- day experimental period.In the morning, feed was weighed.ln the last feeding, excess feed was re- weighed.Feed intake forthe day was determined with reducing weighed feed in the morning and reweighed excess feed in the late afternoon.A feeding tray placed in the center of the cage was used to facilitate feed intake Experimental Habitat and Set-up Three earthen ponds of 200m2 each were used in this experiment.The ponds were dried for a week and limed using Ca (OH), at a rate of 625 kg/ha.Nine net cages of 5.25 x 3.5 x 1 .75m (32 m3 or 18 m2 surface area) each (2.5 cm open mesh size) were suspended in three ponds about 15 cm above the pond boltom.Three cages were placed in each pond.The distance between cages was 1 m.Thereafter, canal water was pumped in the ponds to reach 1 m depth.This water level was maintained during experimental period.
Ponds were fertilized using urea containing 45% ni- trogen (N), TSP containing 21% phosphorous (P) and dry chicken manure containing 1.7% N and 2.8% P at a rate of 3.0 kg N, 1.5 kg P and 10 kg/ha/day, respectively, every three daYs.

Experimental Fish
Silver barb fingerlings of 3-5 cm were purchased from a hatchery.Fish were nursed in a pond and fed on commercial feed containing 30% protein ad libitumfor about a month to attain fingerling fish size of 8-12 cm.Fish were hand-graded to select relatively equal fish size of about 12 cm and then acclimatized to experimental conditions for a week.Thereafter, fish were randomly distributed among cages at a rate of 36 fish/cage.At stocking, individualfish length and weightwere measured, and samples of fish taken from a batch were sacrificed for proximate analysis of their carcasses.Dead fish during the first week of culture occurred as a result of handling and were replaced with equal fish size.Dead fish was counted daily.

Fish Sampling and Harvest
Fish sampling was carried out every 10 days.Fish were counted and weighed for biomass.In the same 100.0Note: (*) CP product; (.*) adopted from NRC (1983): 7.5% MnSO4H2O',3Vo   ZnSOo TH,O, 7% CuSOo.SHrO; 0.6% FeSOo.THrO; 5% NaCl; 7.5% KlO3, g Sy. baH-FrO 42H2O Filler wads added in the formulation to complete 100%.time, dissolved oxygen concentration, pH and water temperature were observed in situ and pond waterwas sampled using a water column water samplerto ob- serve ammonia-N and chlorophyll-a.At harvest, individual length and weight of fish were measured and samples of fish taken from each cage were sacrificed for proximate analysis.

Data Analysis
One-way ANOVA and Least Significant Difference test (LSD) were used to test the statistical signifi- cance between treatments for indices of grovvth, survival, yield, feed and protein intake and utilization.
Treatment means are given + standard error (SE.) in parentheses.Regression analysis was performed to determ ine correlation between variables.

RESULTS
Observed water q uality parameters were comparable forall experimental ponds ffable 2).Survivalwas comparable for all treatments.Dead fish caused by water quality problems were not observed.Reduction in fish number is probably related to handling effects or/and bird predation.

DISCUSSION
GroMh of fish is highly related to availabiirty of proper food.When food is insufficient foi both nraintenance and groMh, groMh wrll be inhibited or cease entirely.Under nonnal condition, silver barb grew at a similar rate with tilapia fingerlings (Milsten et al ,   1995).High groMh rate of silver barb in the present study may be related to the ability of silver barb to utilize natural food efficiently particularly at young stages (Santos,1 993).Possible high contribution of naturalfood to overall nutrient requirement rn ihis ex- periment can be seen from both low daily feed intake and feed conversion ratio.In contrast, Clark ef a/. (1990)reported that high growth rate of red tilapia was related to high feeding rate and available rnicroalgae in a pond environment.
At low dietary proiein levels, available prntein was not enough io support inaxitnurn groMh rate" The avail- able protein rflay be used for energy ratherthan flesh.Wee & l{gamsnae (1987) stated that the slow groMh rate of silver barb at low protein diet is related to the high starch content used in feed formulation.According to fluisnran (1976), available metaLtolit; energy in a diet decreased incretlibly with increasing daily feed intake because of deci'easing feed digestibllity by fish (l-lenken ef a/,, 1986).Excess protein supply was probably observeo at the dietary protein level of 30% as there lvas a tendency for the growth rate to staft to decr^ease (Zeitun et al., 1976; Jauncey, 1 982).
A number of studies showed that the growth iate was posiiively correlated with the dietary protein re- quirernent in g/kgBWday flacon & Cowey. 1985).The present study also showed that growth rate of silver barb increased with increasing protein consumption.
Daily protein intake during rapid growth of silver barb in the present study was aboul 79 protein/kgBWday.This value was lower than that required for common carp fingerlings to attain their maximum growth (about 12 g/kgBWday) (Kausl'rik, 1995).
Evaluation of dietary protein level can be viewed from feed and protein utilization.FCR and PER are kntlwrr to decrease with increasing dietary protein levels (Mazid eta|.,1979, Jauncey,1982;Milsten ef a/., 1995).The present study indicated that FCR significantly decreased with increasing dietary protein con- tent, srnrilarto a trend observed in silver barb (Wee &   Ngamsnae, 1987) and tilapia (Jauncey, 1982).The best FCR (1 .2) was achieved at dietary protein levels above 25%" Hovrever, PER in this study was comparable for fish fed various dietary protein levels.This implies that natural food may contribute in supplying some pro- tein.This finding is comparable for tilapia (Mazid et al , 1 979;Jauncey, 1982;Milsten et al., 1995).How- ever, PEF.( for silver barb in this study was lower than that obserued by Wee & I'Jganrsnae (1987).
Fish carcass cornpositiori at harvest rndicated that there was no difference in the body protein content among fish fed with dift'erent dietary protein levels" This finding was similar to the previous studies of silver bafu arrd tilapia showing that body protein content was not greaily affected by changing dietary protein level (tvlazid et al., '1979; Winfree & Stickriey, 1981;   .Jauncey, 1C82).However, body fat content was lowest at the highest dietary protein level.Reduction in fat carcass content indicates that diets below 30% protern level had excess enerEy.Since this experi- ment used isocaloric feed with different protein levels to maintain similar gross energy (about 4.3 Kcal), the low protein diet contained high carbohydrate (cassava starch) as compared to the higher protein diets.Increase of body fat at lower protein content correlates with high carbohydrate content in supplementary feed (Kaushik, 1995).Zettler et al. (1983) observed decreasing body fat content in line with increasing dietary pro- tein levels for common carp.
Finally, this study has demonstrated that silver barb cultured in a highly ferlilized pond at a rate of ?fish/ m2 performed best on a diet containing 25% protein.
To further reduce feed cost, the rrext study should determine vitamin and mineral content in supplementary feed and a maxinrum standing crop for silver barb in fertilized ponds.
Table1.Composition (%) of experimental feed for silver barb used in Researcher at Research Institute for Openwater Fisheries, JL.Beringin No. 308 Mariana P.O.Box 1125 Palembang ')