Nutritional research in Australia to improve pelleted diets for grow-out barramundi Lates calcarifer (Bloch)

Kevin C Williams[1] and Chris Barlow[2]

Abstract

The farming of Asian seabass or barramundi, Lates calcarifer, is an emerging aquaculture industry in Australia with expected production to exceed 500 t in 1995/96. In Australia, barramundi are fed exclusively on extruded dry diets. Since 1992, an intense research program supported by the Australian Fisheries Research and Development Corporation has examined the nutritional requirements of grow-out barramundi and assessed the nutritive value of locally available protein meals. Optimal feeding practices have been defined for juvenile barramundi held at water temperatures varying incrementally from 20 to 29°C (the range normally experienced on Australian farms). Evaluation of alternative feed ingredients has shown that animal by-product meals such as meat meal and poultry offal meal are highly palatable to barramundi and as well-digested as fishmeal. Vegetable protein meals such as soybean, canola and lupin are less well- digested and not well-liked but can be used cost-effectively for the partial replacement of fishmeal. Increasing the dietary concentration of a reference protein incrementally from 29 to 57% crude protein (CP) caused food intake and food conversion to decrease and improve curvilinearly, respectively, such that growth rate exhibited a bent-stick response, increasing linearly up to about 46% CP. From these studies, the optimum protein to digestible energy (DE) ratio of the diet was estimated to be about 24-25 mg CP/kJDE. The essential fatty acid requirements (as the sum of eicosapentaenoic and docosahexaenoic acids) were found to vary with water temperature from ≈5 mg/g at 20°C to 18 mg/g at 29°C. Under laboratory and commercial farm conditions, diets formulated entirely from terrestrial feed ingredients (except for a low inclusion of fish oil to provide essential n-3 fatty acids) have resulted in as good, if not better, barramundi productivity as conventional diets based on fishmeal. Using trained taste panels, the eating quality of the fish reared on these nil-fishmeal diets has been the same as for conventional diets.

Introduction

In Australia, Asian seabass or barramundi, Lates calcarifer, is a highly-priced recreational and capture fishery. An emerging aquaculture industry is expected to produce more than 500 t of fish in 1995/96 worth AUD $5 M (Figure 1). Most farmed barramundi are sold as plate size (400 to 500 g) whole fish destined for the restaurant trade although there is some interest in growing fish to a larger size (2 to 3 kg) for the fillet trade.

A major impediment to continued expansion of barramundi farming is the high cost of feeding since food comprises 40-50% of on-farm costs. In Australia, all farmed barramundi are grown-out on pelleted (extruded) dry diets which are expensive (AUD $1,200 to $1,500/t). Feed cost is high as diets currently contain large mounts of expensive,imported fishmeal and because of lack of information on the fish's nutrient requirements hinders the development of cost-effective feeds and feeding strategies.

Research to define the nutrient requirements of grow-out fish and shrimp and to assess the suitability of terrestrial protein sources, as cheaper alternatives to fishmeal, is a major priority for Australian aquaculture. This is being addressed in a nationally co-ordinated research program administered by the Australian Fisheries Research and Development Corporation. A large team of aquaculturists from Commonwealth, State and University research institutions and private industry is working collaboratively to develop improved and more cost-effective grow-out diets for barramundi, shrimp (Penaeus monodon), silver perch (Bidyanus bidyanus), and Atlantic salmon (Salmo salar). This paper reviews our work with barramundi to determine their requirements for critically important nutrients and to assess the suitability of locally available terrestrial feedstuffs as cheaper alternatives to fishmeal in manufactured diets.

Effect of water temperature on food intake and growth

In Australia, barramundi are grown-out typically in cages suspended in estuarine water or in fresh-brackish water in earthen ponds. In areas where barramundi are farmed, water temperature varies seasonally between 20°C and 29-30°C. Because water temperature is known to have a profound effect on food intake of aquatic animals (Braaten, 1978; Steffens, 1989; Talbot, 1993), studies to define optimal feeding practices for juvenile (≈30 to 300 g) barramundi examined the effects of water temperature, feeding frequency and fish size (weight). Food intake (of a dry pellet containing: dry matter, 95%; crude protein, 44%; and estimated digestible energy, 15 kJ/g) of acclimatised fish increased essentially linearly with water temperature (over the range of 20 to 29°C) and fish size (Figure 2A); expressed as a function of fish biomass, food intake declined allometrically with fish size (Figure 2B).

Absolute growth rate increased linearly with fish size at each water temperature (Figure 3).

Increasing the feeding frequency from 1 to 3 times daily increased food intake, slightly especially for smaller fish

Assessment of nutritive value of feed ingredients

Measurement of the apparent digestibility of a feedstuff is essential if diets are to be formulated to meet prescribed nutrient specifications at least cost. Because of the difficulty if not impossibility of collecting the total daily faecal output of an aquatic animal, apparent digestibility is typically measured using indirect procedures employing digestibility markers. Differences in the concentrations of the marker and of the particular nutrient in the food and in representative samples of faeces allows digestibility to be derived from the equation: (<100 g), but the extra food did not result in better growth rate. A similar observation had earlier been made by Tucker et al. (1988). Analysis of the data generated the following food intake prediction equation:

lnDFI = -7.285 + 0.478lnW + 0.391T -0.0065T² + 0.074F (R² = 0.97)

where ln is the natural logarithm, DFI is daily food intake (g/fish/d), W is weight (g) of the fish, T is water temperature (°C) and F, the number of feeds/d.

AD_Nut = 100 * [ 1 - {(M_FI/M_FO) * (Nut_FO/Nut_FI)}]

where AD is apparent digestibility (%); M and Nut are the concentrations (% dry matter) of the marker and nutrient respectively in the food (FI) and faeces (FO).

Figure 1. Production of barramundi from capture fisheries and aquaculture in Australia

We have found Ytterbium acetate (at 0.05 to 0.1% of diet) to be a more reliable digestibility marker than chromic oxide. Its analysis however, requires a mass spectrophotometer. Apparent digestibility measurements were made using substitution procedures with the test ingredient being substituted in a basal diet at amounts of not less than 30%. We found faecal samples collected by sedimentation resulted in an over- and under-estimation of the apparent digestibility of protein and lipid respectively, because of the rapid leaching of soluble N which was almosthalf of the total N in the faeces (Windell et al., 1978; Smith et al., 1980; Williams et al., 1996). While intestinal dissection is the sacrificed and the procedure is very labour intensive. However, stripping of lightly anaesthetised fish has proved to be a reliable and efficient method for collecting faecal samples. From the the data on crude protein and energy apparent digestibility values for a number of dry feed ingredients commonly available in Australia (Table 1), barramundi are capable of digesting the protein from a wide variety of animal and plant feedstuffs but that they are less well able to digest the energy contained in terrestrial animal and plant food sources.

Protein requirement of juvenile barramundi

Many different approaches have been used with terrestrial and aquatic animals to define essential amino acid requirements. The most widely used (traditional) methodology involves feeding graded levels of one amino acid at a time in a test diet containing either all crystalline amino acids or a mixture of pure proteinsand crystalline amino acids. Disadvantages of this methodology are (i) it is a slow process to evaluate each of the 10 or so essential amino acids; (ii) absolute response to diets comprised mostly of crystalline amino acids is usually inferior to that seen with diets based on intact proteins; and (iii) the derived dietary amino acid level that maximises fish response will be specific to the experimental conditions, particularly energy intake and the adequacy of all of the other essential amino acids.

Figure 2. Effect of water temperature and fish size on intake of dry food pellet

by juvenile barramundi: A, daily food intake; B, percent of fish biomass

Figure 3. Effect of water temperature and fish size on growth rate of barramundi

Table 1. The apparent digestibility of air-dry feed ingredients for barramundi

Feed ingredient	Apparent digestibility (%)		Digestible energy
	Crude protein	Gross energy	(kJ/g)
Fishmeal (Danish)^1	88.7	99.2	20.0
Fishmeal (tuna)^1	92.3	68.1	11.2
Meat meal (55% CP)^2	75.1	76.3	13.4
Meat meal (50% CP)^2	60.4	63.5	12.0
Poultry offal meal^2	75.8	73.6	15.8
Soybean meal (full-fat)^2	82.3	72.2	15.7
Soybean meal (solv)^2	80.8	59.3	11.8
Canola meal^2	80.0	54.2	10.7
sem (range)	0.9-10.2	1.9-8.4	0.4-1.8

¹ Determined by intestinal dissection (Williams et al., 1996).

2 Determined by stripping of fish (McMeniman et al., 1996)

An alternative methodology which has gained considerable support over the last decade is the "ideal protein" concept as espoused by Cole (1980). 'Ideal' protein is defined as one that is perfectly balanced in terms of its amino acid content for the type of production required (viz for growth, maintenance, reproduction). Such a protein would have the highest possible biological value, i.e., the greatest efficiency of conversion of dietary protein into deposited protein. Once this is determined, dietary specifications can easily be tailored for any given rate of growth (strictly speaking, for a given rate of protein deposition) of the fish. If growth in fish mirrors that seen in terrestrial monogastric animals such as pigs and poultry, growth is expected to exhibit dependency and independency to both protein and energy intake. The optimum dietary protein to energy ratio can be determined by feeding increasing amounts of protein (of constant quality) in conjunction with aconstant amount of energy. The slope of the response line (i.e., biological value) will indicate how close the amino acid composition of this protein is to the ideal pattern.

This approach was tested by formulating a semi-purified diet in which all of the protein (of an amino acid composition closely matching that of barramundi protein) was provided as a protein mixture (reference protein). Protein content of the diet was varied incrementally from 29 to 57% by adding the reference protein at the expense of non-protein ingredients manipulated to maintain the desired energy content (Table 2). Fish were fed to satiety twice daily and held in water at 28°C for an experimental period of 28 d. Production responses are tabulated in Table 3 and Figure 4. As the inclusion content of the reference protein increased, there was a marked curvilinear reduction in food intake and a corresponding although less marked improvement in food conversion (P<0.05). These effects caused growth rate to exhibit a bent stick response, increasing linearly up to a dietary protein content of about 46%. This is similar to the recommendation of Boonyaratpalin (1989) that the dietary crude protein content of grow-out Asian sea bass should be 45 to 50% (supplied predominantly from fishmeal).

When expressed as a function of absolute intake, growth and food conversion improved curvilinearly (P<0.05) with increasing protein intake with the response reaching an asymptote value at an intake of 1.44 g protein/fish/d (Figure 5A). In contrast, growth rate and food conversion deteriorated with increasing digestible energy consumption (Figure 5B), indicating that the response was clearly that of a simple protein dependency. Based on this result, the dietary protein to digestible energy dependency.

Table 2. Composition of the diets in the protein requirement experiment

Feed source	Diet
	1	2	3	4	5	6
	Formulation (%)
Reference protein^1 Starch (autoclaved) Diatomaceous earth Soybean oil Fish oil Vit + Min premix	35.0 47.5 3.5 2.0 8.0 4.0	42.0 40.0 5.3 1.8 7.2 4.0	49.0 32.5 7.1 1.6 6.4 4.0	56.0 25.0 8.9 1.4 5.6 4.0	63.0 17.5 10.7 1.2 4.8 4.0	70.0 10.0 12.5 1.0 4.0 4.0
	Chemical analysis
Crude protein (%) Gross energy (kJ/g) Est dig. energy (kJ/g) Crude fat (%)	29.0 18.93 15.0 11.7	34.6 18.85 15.0 11.0	40.1 18.78 15.0 10.2	45.7 18.70 15.0 9.5	51.2 18.63 15.0 8.7	56.8 18.55 15.0 8.0

¹ Formulation (g/kg) of the reference protein was: Casein, 430; Fishmeal (Peruvian), 300; Gluten, 250; Lysine HCl, 5; d/l Methionine, 5.5; l Threonine, 2.5; l Tryptophan, 1; and NaHCO₃, 6.

Table 3. Production responses of barramundi to diets varying in protein content

Attribute	Treatment (diet CP%)						±sem
	29.0	34.6	40.1	45.7	51.2	56.8
Start weight(g) End weight (g)	74.9 123.6^C	75.2 144.2^B	78.6 153.5^A	76.3 157.5^A	75.8 152.1^A	76.2 158.5^A	1.47 2.14
Food intake (g/d)	4.16^A	3.62^B	3.32^C	3.00^D	2.61^E	2.66^E	0.058
Growth (g/d)	2.38^C	2.51^BC	2.72^AB	2.90^A	2.72^AB	2.94^A	0.072
FCR (g:g)	1.76^E	1.44^D	1.22^C	1.04^B	0.96^AB	0.91^A	0.026

^A,B,C,.D - Means without a common superscript letter differ (P<0.05).

Table 4. Effect of water temperature (WT) and essential fatty acid content (EFA) of the diet on the growth performance of juvenile barramundi

WT	EFA( EPA + DHA) content (mg/g)						Response^1	±sem
(°C)	5.1	8.3	11.5	14.6	17.8	21.0	(WT x EFA)
	Food intake (g/d)
20	1.29	1.31	1.34	1.19	1.19	1.25	ns
29	4.41	4.21	3.89	3.58	3.69	3.93	L; Q	0.091
	Growth rate (g/d)
20	0.70	0.72	0.81	0.70	0.79	0.75	ns
29	2.93	3.07	3.16	3.13	3.44	3.24	L; Q	0.103
	Food conversion (g:g)
20	1.83	1.83	1.66	1.70	1.52	1.68	Q
29	1.50	1.37	1.23	1.14	1.07	1.18	Q	0.077

1 Response to diet at each water temperature: ns, not significant (P>0.05); L, Linear (P<0.05); Q, Quadratic (P<0.05).

At low water temperature, fish response was unaffected by dietary fatty acid content, whereas at high water temperature, food intake declined curvilinearly with increasing n-3 fatty acid content (Figure 6). This is because of a concomitant improvement in food conversion, growth rate improved linearly up to a total EPA and DHA content of 17.8 mg/g (Figure 7).

The observed curvilinear response of food intake to dietary fatty acid content at high water temperature was an unexpected result. The increased food intake on the diets containing the lowest n-3 fatty acid content could be interpreted as an attempt by the fish to increase its intake of critical n-3 fatty acids. This is plausible since food conversion also showed a marked deterioration for diets containing the lowest n-3 fatty acid content. The lack of response to dietary fatty acid content by fish held at low water temperature was probably due to the reduced growth and thus a minimal requirement for n-3 fatty acids. These results indicate that the optimal dietary n-3 to n-6 fatty acid content should be not less than 1.6:1 (equivalent to an EPA + DHA content of 17.8 mg/g) for rapidly growing fish at high water temperature whereas at low water temperature the ratio need not be greater than 0.6:1 (EPA + DHA content of 5 mg/g). In reviewing the essential fatty acid requirements of marine fishes, Tucker (1992) concluded that a dietary EPA + DHA concentration of 20 mg/g was a reasonable specification for the young of most species but this could be reduced to 14 mg/g for older fish. Tucker (1992) stressed the essentiality of DHA and advocated that it comprise at least half of the n-3 fatty acid content of the diet. Boonyaratpalin (1989) recommended that the total n-3 fatty acid content of the diet for juvenile Asian sea bass should be 10 to 15 mg/g.

Figure 6. Effect of dietary fatty acid content and water temperature on food intake of juvenile barramundi

Figure 7. Effect of fatty acid content of the diet and water temperature on growth rate of Juvenile barramundi

Commercial trailing of nil-fishmeal grow-out diets for barramundi

The primary objective of the research program was to develop improved and cheaper barramundi grow-out diets with a reduced dependency on fishmeal. Information on the nutritive value of alternative feedstuffs and the fish's requirements for key nutrients was used to formulate practical diets for commercial evaluation. Several laboratory and on-farm trials have been done to demonstrate the suitability of these new generation diets. The results of a study comparing diets formulated without any fishmeal with either a proprietary barramundi diet or a high fishmeal experimental control diet are discussed to illustrate the progress that has been made.

A 4x4 randomised block design was used to compare three experimental diets (two containing no fishmeal) with a proprietary barramundi diet, all being commercially extruded as dry floating pellets (Table 5). The ingredient cost of the nil-fishmeal diets was 15 to 20% cheaper than that of the proprietary diet.

Cages (2m²⁾ were stocked with 300 fish (initially 226 ±16.3 g) and suspended in an aerated freshwater pond. Fish were fed once daily to satiety and reared on the diets for 10 weeks. At the conclusion of the feeding period, all fish were weighed and samples taken to assess eating quality using taste-panel procedures.

There were significant (P<0.05) differences in fish growth performance between the diets (Table 6). Food intake of fish on both of the nil-fishmeal diets (diets 2 and 3) was higher than on each of the other diets, indicating high acceptability by the fish for the nil-fishmeal diets. Food conversion and growth rate on the high energy nil-fishmeal diet (diet 2) were as good if not better (P<0.05) than all of the other diets. Food conversion was best on the fishmeal control diet (diet 1) but not significantly better (P>0.05)than that for the high energy nil-fishmeal diet).

These results demonstrate that appropriately formulated and cheaper diets without fishmeal (but containing some fish oil as a source of n-3 fattyacids) are able to grow barramundi as well as those fed on conventional high fishmeal diets. Equally important, the eating quality of the fish reared on nil- conventional high fishmeal diets. Equally important, the eating quality of the fish reared on nil-fishmeal diets was indistinguishable from fish fed on high fishmeal diets.

Assessment of the eating quality of the fish using trained taste panels at the Queensland Government's Centre for Food Technology showed similar scores for all diets (Table 7). Scores for undesirable off-colours and flavours were very low and the overall acceptance of the fish on all diets was very high. than that for the high.

Further studies are continuing to specify requirements of grow-out barramundi for critical essential amino acids and the role of high energy diets for the commercial production of the fish.

The financial support of the Australian Fisheries Research and Development Corporation and the Australian Meat Research Council is gratefully acknowledged. We thank the Workshop Organising Committee and in particular Dr. Michael Phillips of NACA and Mr. Rooney Biusing of the Sabah Department of Fisheries for the invitation to present this work.

Table 5. Composition of the diets fed in the on-farm trial

Attribute	Diet description and formulation
	Diet 1 (Control)	Diet 2	Diet 3		Diet 4(Proprietary^1)
	Formulation (g/kg)
Wheat	304	105	161
Chile fishmeal (65% CP)	350	0	0
Meat meal (52% CP)	0	500	500
Meat meal (60% CP)	100	0	0
Blood meal (ring)	0	90	70
Soybean (fullfat)	160	100	150
Gluten (90% CP)	50	100	50
l-lysine HCl	0	7.5	6.5
d/l Methionine	1.5	3	3
Fish oil (Chile)	25	60	50
Tallow	0	25	0
Vit & min mixture	9.5	9.5	9.5
	Chemical analysis
Gross energy (kJ/g)	19.2	21.0	19.6	20.0
Est DE (kJ/g)	15.0	16.4	15.3	nd
CP (g/kg)	436	470	440	543
Fat (g/kg)	87	138	116	69
Arginine (g/kg)	27.4	29.4	28.8	29.7
Lysine (g/kg)	28.7	30.2	28.5	46.1
Meth + Cyst (g/kg)	9.2	8.1	7.3	10.8
Threonine (g/kg)	17.4	16.0	15.2	23.9
C20:5 n-3 (g/kg)	4.3	6.7	5.5	5.0
C22:6 n-3 (g/kg)	7.5	8.9	7.4	9.1

¹ Formulation of the proprietary diet is confidential. nd, not determined.

Table 6. Effect of diet on performance of barramundi reared under commercial farm conditions

Response attribute	Diets				±sem
	Diet 1	Diet 2	Diet 3	Diet 4
Food supply (kg/wk/cage)	7.6^C	9.1^A	9.1^A	8.3^B	0.10
Growth rate (kg/wk/cage)	6.2^B	7.0^A	6.4^AB	6.1^B	0.18
Farm food conversion	1.22^A	1.31^AB	1.44^B	1.37^B	0.041

^A,B within response attributes, means without a common letter differ (P<0.05)

Table 7. Effect of diet on eating quality scores (0 = low; 100 = high) of fish reared under commercial farm conditions

Response attribute^1	Diets
	Diet 1	Diet 2	Diet 3	Diet 4
Colour Yellow Grey	6.9 10.5	8.8 10.5	9.1 9.7	7.6 9.5
Flavour Sweet Fishy Muddy	19.2 49.0 14.8	22.4 47.3 13.9	21.7 45.5 15.9	18.1 46.8 16.6
Texture Firm Moist	46.5 44.2	46.9 43.9	44.3 42.7	47.3 47.4
Overall acceptability	60.0	64.3	61.2	63.5

¹ Differences between diets for all attributes were not significant (P>0.05).

References

Barlow, C., K. Williams, L. Rodgers, C. Agcopra, I. Hocking. 1996. Effect of water temperature and dietary w-3 to w-6 fatty acid ratio on growth of juvenile Asian sea bass, Lates calcarifer (Bloch). pp. 29-30. In: Proc. World Aqua. Soc. 29 January-2 February 1996, Bangkok.

Boonyaratpalin, M. 1989. Seabass culture and nutrition. pp. 43-77. In: Akiyama, D. M. (ed.) 'Proceedings of the People's Republic of China Aquaculture and Feed Workshop'. Singapore: American Soybean Association.

Braaten, B. R. 1978. Bioenergetics - A review on methodology. pp. 462-501. In: Proc. World. Symp. Finfish Nutr. Fishfeed Tech. Hamburg, 20-23 June 1978.

Cole, D. J. A. 1980. The amino acid requirements of pigs - the concept of an ideal protein. Pig News Info. 1:201-205.

McMeniman, N. 1996. In: Allen, G.L. (ed.) FRDC Progress Report - Fishmeal Replacement in Aquaculture Feeds for Barramundi, April 1996. Canberra, Australia: Fisheries Research & Development Corporation.

Smith, R.R., M.C. Peterson, and A.C. Allred. 1980. Effect of leaching on the apparent digestion coefficients of feedstuffs for salmonids. Prog. Fish Cult. 42: 145.

Steffens, W. 1989. Energy requirement. pp. 184-208. In: Principles of Fish Nutrition. Chichester, UK: Ellis Horwood Ltd.

Talbot, C. 1993. Some aspects of the biology of feeding and growth in fish. Proc. Nutr. Soc. 52: 403-416.

Tucker Jr. , J. W. 1992. Marine fish nutrition. pp. 25-40. In: Allen, G. L. and W. Dall (eds.). Aquaculture Nutrition Workshop. New South Wales Fisheries, Salamander Bay, Australia.

Tucker Jr., J. W., M. R. MacKinnon, D. J. Russell, J. J. O'Brien, and E. Cazzola. 1988. Growth of juvenile barramundi (Lates calcarifer) on dry feeds. Prog. Fish Cult. 50: 81-85.

Williams, K., C. Barlow, J. Rose, B. Kelly. 1996. Effect of faecal collection method on the apparent digestibility of diets for Asian sea bass, Lates calcarifer (Bloch). pp. 438-439. In: Proc. World Aqua. Soc. 29 January-2 February 1996, Bangkok.

Windell, J.R., J. W. Foltz, and J. A. Saroken. 1978. Methods of faecal collection and nutrient leaching in digestibility studies. Prog. Fish Cult. 40: 51-55.

---

[1] CSIRO Division of Marine Research, P O Box 120, Cleveland Qld. 4063, Australia

[2] QDPI Freshwater Fisheries and Aquaculture Centre, Walkamin Qld. 4872, Australia