DIGESTIBILITY OF SELECTED FEED INGREDIENTS FOR TIGER GROUPER , Epinephelus fuscoguttatus

The apparent digestibility (AD) of eight feed ingredients are widely available in Indonesia was determined. In each of two 5x5 latin-square experimental, tiger grouper Epinephelus fuscoguttatus juveniles (100—150 g) were fed a reference diet and four test diets in accordance with the latin-square design. Test feed ingredients were substituted at rates of 40% for animal meals or 30% for plant meals. Chromic oxide was used as the digestibility marker. In determining the ingredient AD, the substitution ratio was calculated as the proportion of the nutrient (or energy) contributed by the test ingredient on an ‘as-is’ basis. Digestibility tanks were steeply slope 200 L cylindroconical tanks with a bottom outlet to facilitate faecal collection, which was carried out at 3-hourly intervals throughout the day. Each collection period took 5—7 days with a similar acclimatization time between diets. A combined ANOVA of the data for both experimental showed no difference (P>0.05) in the AD’s for each reference diets. Thus for comparative purpose, the derived AD’s of the test ingredients were analysed as a single ANOVA. The digestibility of animal meals was generally high (>59% for dry matter, >83% for protein, >65% for lipid, and >70 for gross energy) while that of plant meals was slow (<53% for dry matter, <53% for protein, <66% for lipid, and <46% for gross energy). This information will enable grow-out feeds for tiger grouper to be formulated on a least-cost digestible nutrient basis.


INTRODUCTION
Marine feed ingredients and specifically fish meal, are in high demand as protein sources for fish feeds.Fish meal is regarded as the best protein source for fish feeds, being high in protein, has a close to ideal profile of essential amino acids and few if any anti-nutritional factors (Allan et al., 2000).The world demand for fish meal keeps increasing.However, world production of fish meal has not increased beyond a production of about 6.5 million to nnes per ye ar since the early 1990's (Shepherd et al., 2005).Tiger grouper require high dietary protein of 47%-50% (Laining et al., 2003a;Giri et al., 2004), and typically manufactured diets contain from 40 to 62% of expensive fish meal.Replacement of fish meal with a cheaper, high-quality source of protein might reduce the production costs of culturing high tropic level species such as tiger grouper, Epinephelus fuscoguttatus.The assessment of different so urce s o f pro te in fo r fish fe ed is an international and increasing research priority (Manzi, 1989;Hardly & Kissil 1997in Allan et al., 2000).
Apart from considerations of palatability and anti-nutritional factors, the apparent digestibility of an ingredient determines how much of the nutrients in the ingredient can be utilized by the animal.The higher the digestibility of the ingredient, more of its nutrient content will be available to support growth and other metabolic functions of the animal.Consequently, the digestibility of an ingredient is an important measure of an ingredient's nutrient quality.
The potential sources of protein available in large quantities in Indonesia are poultry offal, golden snail, green mussel and mysid meal.Yellow and white corn meal, rice bran, and sorghum meal are also carbohydrate sources that are in plentiful supply in Indonesia.This paper reports a digestibility study to determine the apparent digestibility of these feed ingredients because of their potential to be used in the manufacture of cost-effective feeds for tiger grouper.

Experiment Overview
The study comprised two independent 5 x 5 latin square experiment with five diets, a reference and four test feed ingredient diets, being examined in each experiment.The refe rence die t was the same for e ach experiment with four test feed ingredients of animal origin-poultry offal meal, mysid meal, golden snail meal and green mussel mealbeing examined in Experiment 1and four of plant origin-fine rice bran, corn meal (yellow and white) and sorghum meal-being examined in Experiment 2. Test feed ingredients were substituted into the reference diet at the rate of either 40% or 30% for ingredients of animal or plant origin, respectively (Table 1).Chromic oxide was added at an inclusion rate of 1% of the diet as the digestibility marker (Takeuchi, 1988;De Silva & Anderson, 1995).
All test feed ingredients were obtained locally.Each set involved five faecal collection periods with all fish, randomly in turn, being fed each of the five diets in accordance with the latin-square design of the experiment.The proximate nutrient and gross energy composition of the test ingredients are shown in Table 2.

Diet t ype
As used product (g)

Ingredient
Table 1.Formulation of diets used in determining the apparent digestibility of feed ingredients in Experiment 1 and 2 One cohort of juvenile tiger grouper, Epinephelus fuscoguttatus, bred at the Gondol RIM, Bali hatchery was air-transported to South Sulawesi and acclimated to the floating net cage facility at Awerange Bay, Barru Regency, South Sulawesi.Prior to starting on the experiment, fish were held for several months in 2x2x2 m floating net cages and acclimatized to being fed a moist reference diet.Afterwards, the fish were selected on the basis of freedom from defects, initial weight of 100-150 g and then distributed into 5 floating net cages of 1x1x2 m at a density of 25 fish/cage.For each collection period, the fish were acclimated to their test diet for 5 d and fed twice daily, morning and late afternoon.On the 6 th day, 20 fish from each cage were carefully transferred into 200 L conical, 35% slope, bottom tank for faecal collection.In each period, faeces were collected over 5-7 days, which enabled a sufficient quantity of faeces to be collected for chemical analysis.

Chemical Analyses and Calculations
A representative sample of feed or dried faeces was homogenized using a mortar and pestle and analyzed essentially by AOAC International (1999) procedures: dry matter (DM) by oven drying at 105ºC for 16 h; ash by ignition in a muffle furnace at 550ºC for 24 h and crude protein by micro-Kjeldahl analysis with distillation into 4% boric acid and titration with sulphuric acid using methyl red indicator for end point determination.Total lipid was dete rmine d gravimetrically fo llo wing chloroform: me thanol extraction o f the sample (Bligh & Dyer, 1959); energy by bomb calorimetry and chromium after acid digestion using a Shimadzu UV-VIS 2401 PC spectrophotometer.
The apparent digestibility coefficient (ADC) of dry matter (DM), crude protein (CP), total lipid (TL) and gross energy (GE) for the reference and test diets were calculated using the equation: Where AD N I , AD NT a nd AD NR a re appa rent digestibility o f the nutrient in the ingredient, the test diet and the reference diet, respectively.The substitution rate (a) was calculated as the proportion of the nutrient (or energy) contributed by the test ingredient in the reference diet on an 'as-fed wet' basis (Bureau et al., 1999).
The ADC data for diets and those derived for the test feed ingredient in each experiment were subjected to a least squares analysis of variance for a latin square design, isolating effect due to column (fish) and row (collection period).A combined analysis of the ADC's for diets for the two experiments found no significant difference between the reference diets in each experiment (P>0.05).Thus, the derived ADC for the test feed ingredients were analysed as a single factor ANOVA with differences between treatments (ingredients) te ste d fo r significance using Fisher's protected t-test (Snedecor & Cochran, 1989) wherein differences between means were examined only when the F-test of the ANOVA was significant (P<0.05).

RESULTS AND DISCUSSION
Digestibility coefficients for the diets fed in experimental 1 and 2 are detailed in Tables 3 and 4, respectively.A statistical analysis of the data for both experiments, isolating the effects of experiment, diet and associated interactions, showed that the observed ADCs for each of the reference diets were not significantly different (P>0.05) from one another.Thus, the ADCs derived for each of the test feed ingredients in each experiment were analysed as a combined ANOVA and the results presented in Table 5.
The apparent digestibility of the protein of animal meals was high (77% to 89%) and not significantly different from each other but significantly higher than for the plant meals (41% to 52%), which were statistically not diffe rent fro m e ach o the r.To tal lipid digestibility tended to follow a similar pattern as for protein digestibility but differed in that golden snail meal was lower than for the other  animal meals (62 cf 84%-94%, respectively) while the rice bran was higher than for the other plant meals (66 cf 24%-31%, respectively).Dry matter and energy digestibilities of poultry offal meal, mysid meal and green mussel meal were not significantly different from each other and significantly higher in the case of the mysid and mussel meals than the golden snail meal, which was significantly higher than the plant meals.For the plant meals, rice bran had the lowest, and sorghum the highest, DM and energy apparent digestibilities.
Tiger grouper digested the animal feed ingredients more efficiently than the plant feed ingredients with DM, CP, total lipid and energy ADC's generally being significantly higher.Of the animal meals, golden snail meal was the least digestible while the fibre-rich rice bran had the lowest DM, CP and energy digestibilities of the plant meals.However, the lipid dige stibility o f the rice bran was comparatively high, much higher than all other plant meals and equivalent to that of golden snail meal.Similar results wherein animal meals have high digestibility and plant meals much lower digestibilities have been reported for other carnivorous fish such as rainbow trout Oncorhynchus mykiss (Austreng et al., 1977), red drum Sciaenops ocellatus (Gaylord & Gatlin, 1996;McGoogan & Reigh, 1996), yellow tail Seriola quinqueradiata (Masumoto et al., 1996) humpback Cromileptes altivelis and gold-spot Epinephelus coioides groupers (Laining et al., 2003b;Eusebio et al., 2004), catfish Clarias spp (Usman et at ., 2003) and cobia Rachycentron canadum (Zhou et al., 2004).The low ADC's of these plant ingredients were probably due to their high content of nitrogen free extract (NFE), particularly the complex carbohydrate components of seeds, which are poorly digested by piscivorous fish (Lupatsch et al., 1997;Cowey & Walton, 1989).
The low protein ADC of the plant meals was not unexpected as studies with other grouper species have shown similar low digestibilities.For instance, tiger grouper digested only about 47% of the protein of rice bran in this study, which was in the range of values of 43% for gold-spot grouper (Eusebio et al., 2004) and 59% for humpback grouper (Laining et al. 2003b).However, higher protein digestibilities for rice bran were reported by Sullivan & Reigh (1995) for hybrid striped bass Morone saxatilis x M. chrysops (71%) and by McGoogan & Reigh (1996) for red drum (77%).Such effects may indicate differences between species in their capacity to digest the fibre matrix of rice bran although differences in how the rice bran was handle and processed may equally have played a role.In our study, the rice bran originated from a local rice mill and was used without any further refining.Importantly, all diets were fed as moist pellets and manufactured using a cold extrusion process.Cooking of the meal, either

Test feed ingredient
that which results from steam conditioning of the mix prior to pelleting or that occurring during hot extrusion processing can increase starch gelatinization and improve digestibility (De Silva & Anderson, 1995;Davis & Arnold, 1995).The rice bran examined in the above cited hybrid striped bass and red drum studies was finely ground prior to mixing and dried after pelleting.This processing may have improved the digestibility of the rice bran.
The digestibility of unsaturated lipids, including those from plants, is usually high for carnivorous fish with estimates typically between 90 and 100% (Caballero et al., 2002;Carter et al., 2003;Usmani et al., 2003;Zhou et al., 2004).In the present study, lipid digestibility was high (>84%) for each of the animal meals other than for golden snail meal, which at 62% was slightly lower than rice bran (66%).The comparatively low lipid digestibility of the snail meal was surprising since about two thirds of its lipid comprises unsaturated fatty acids and of these one third are fatty acids with three or more unsaturated bonds (Bombeo-Tuburant et al., 1995).It is well estabilished that the digestibility of lipid in fish, and animals generally, is influenced by chain length and saturation of the constituent fatty acids, decre asing with incre asing le ngth and incre asing with incre asing de gre e o f unsaturation (Olsen et al., 1998;Johnsen et al., 2000;Menoyo et al., 2003;Ng et al., 2004).Perhaps the quantitatively low content of lipid in the snail meal (4.9%) might explain the apparent anomalous result although the comparatively low DM and protein digestibility values for snail meals suggests that other factors may also have had a role in reducing overall digestibility.The low lipid digestibility of the cereal grains (24% to 31%) is also an anomalous result but probably due to their low lipid contents (1% to 5%) and the associated mathematical errors in deriving the estimate rather than the lipid in these meals being poorly digestible.
There was a clear distinction between the animal and plant meals in the amount of digestible energy they provided: 71% to 83% for the animal meals and 39% to 45% for the plant meals.This was to be expected given the low protein, DM and lipid digestibility of the plant meals and the comparatively high digestibility of the animal meals.A similar result has been reported for other carnivorous fish (Sullivan & Reigh, 1995;Burel et al., 2000;Booth et al., 2005).
In conclusion, the study has shown that tiger grouper efficiently digest animal feed ingredients, indicating that the ingredients have po te ntial to be use d as die tary replacements for fish meal.Some caution is advised for golden snail meal since its overall digestibility was inexplicably poor, perhaps indicating that unknown factors may be affecting its nutritional value.While the plant meals were not easily digested, nevertheless they are an integral part of compounded diets and data on their digestibility are important when formulating diets to satisfy the animal's re quireme nts for dige stible nutrie nts.However, steam conditioning of these meals prior to pelleting or using hot extrusion manufacturing procedures may increase their digestibility for tiger grouper.The high digestibility of green mussel suggests its polyculture with fish may provide an additional local source of protein for the fish while helping to alleviate nutrient output from farms.Until more eco-friendly pelleted feeds become comme rcially available, such integrated systems may be an effective way of reducing the e nvironme ntal impacts o f cage aquaculture.

Table 2 .
The proximate nutrient and gross energy composition of air-dry test feed ingredients

Table 3 .
The dry matter (DM), crude protein (CP), total lipid (TL) and gross energy (GE) apparent digestibility coefficients of the diets examined in Experiment 1 A,B,C,D Within columns, means without a common letter differ (P<0.05)

Table 4
. The dry matter (DM), crude protein (CP), total lipid (L) and gross energy (GE) apparent digestibility coefficients of diets examined in Experiment 2

Table 5 .
The dry matter (DM), crude protein (CP), total lipid (TL) and gross energy (GE) apparent digestibility coefficients of the test feed ingredients derived from diets examined in Experiment 1 and 2