Fish nutrition - An Overview
A few notes about the Fish Nutrition Research Lab by Dominique Bureau ( firstname.lastname@example.org )
The Fish Nutrition Research Laboratory is a joint venture between the University of Guelph and Ontario Ministry of Natural Resources. It is a research facility in existence since 1969. It conducts basic and applied research in the field of fish nutrition and feeding, with particular emphasis on nutrient utilization and requirements, bioenergetics, digestibility, feed formulation, feeding systems and waste management. The studies carry out in the lab are mostly with salmonids and supported by the Ontario Ministry of Natural Resources and the Ontario Ministry of Agriculture, Food and Rural Affairs, the Department of Fisheries and Oceans, AquaNet, NSERC, and the industry. Visitors welcome.
In culturing fish in captivity, nothing is more important than sound nutrition and adequate feeding. If the feed is not consumed by the fish or if the fish are unable to utilize the feed because of some nutrient deficiency, then there will be no growth. An undernourished animal cannot maintain its health and be productive, regardless of the quality of its environment.
The production of nutritionally balanced diets for fish requires efforts in research, quality control, and biological evaluation. Faulty nutrition obviously impairs fish productivity and results in a deterioration of health until recognisable diseases ensues. The borderlines between reduced growth and diminished health, on the one hand, and overt disease, on the other, are very difficult to define. There is no doubt that as our knowledge advances, the nature of the departures from normality will be more easily explained and corrected. However, the problem of recognizing a deterioration of performance in its initial stages and taking corrective action will remain an essential part of the skill of the fish culturist.
Protein is required in the diet to provide indispensable amino acids and nitrogen for synthesis of non-indispensable amino acids. Protein in body tissues incorporate about 23 amino acids and among these, 10 amino acids must be supplied in the diet since fish cannot synthesise them. Amino acids are need for maintenance, growth, reproduction and repletion of tissues. A large proportion of the amino acid consumed by a fish are catabolized for energy and fish are well-adapted to using an excess protein this way. Catabolism of protein leads to the release of ammonia.
Protein is the most important component of the diet of fish because protein intake generally determines growth (protein growth has, in general, priority), has a high cost per unit and high levels are required per unit of feeds.
First observations on fish protein and amino acid requirements came from studies on natural diet of different fish. Natural diet (plankton, invertebrates, fish) is generally rich in protein and has a good amino acid balance. All dietary proteins are not identical in their nutritive value. The nutritional value of a protein source is a function of its digestibility and amino acid makeup. A deficiency of indispensable amino acid creates poor utilization of dietary protein and hence growth retardation, poor live weight gain, and feed efficiency. In sever cases, deficiency reduces the ability to resist diseases and lowers the effectiveness of the immune response mechanism. For example, experiments have shown that tryptophan-deficient fish become scoliotic, showing curvature of the spine, and methionine deficiency produces lens cataracts. Salmonid diets generally contain 35-45% digestible protein (DP), or 40-50% crude protein. However, amino acids or protein must be supplied in relation to digestible energy (DE). The recommended ratio of protein to energy in the salmonid diet is 20-26 g DP/MJ DE (92-102 g protein per Mcal). Increasing these proportions increases ammonia excretion; the requirement for dissolved oxygen is also increased because the efficiency with which the energy is used is decreased.
Why do fish have such high requirements for protein? The main factors explain this phenomenon:
1) The protein requirement in terms of dietary concentration (% of diet) is high but the absolute requirement isn’t (g/kg body weight gain). This is due to the fact that fish have a lower absolute energy requirement than mammals. This results in similar g body weight gain/g protein ingested as mammal but better feed efficiency (gain:feed).
2) Protein (amino acids) is used as a major energy source. Some economy can be made here if other dietary fuel are present in adequate amounts, e.g. increasing the lipid (fat) content of diet can help reduce dietary protein (amino acid) catabolism and requirement. This is referred to as protein-sparing effect of lipids. Protein to useful energy ratio is the factor that should be considered, not % protein of the diet per se.
10 Indispensable amino acids
Histidine (His) Isoleucine (Iso) Leucine (Leu)
Table 1. Indispensable amino acid requirements of different species of teleost (g / 100 g protein)
Table 2. Amino acid composition of common protein sources (g/ 100 g protein).
Lipids (fats) encompass a large variety of compounds. Lipids have many roles: energy supply, structure, precursors to many reactive substances, etc. In the diet or carcass of fish, lipids are most commonly found as triglycerides, phospholipids and, sometimes, wax esters. Triglycerides are composed of a glycerol molecule to which three fatty acids are attached. Phospholipids are also composed of a glycerol molecule but with only two fatty acids. Instead of a third fatty acid a phosphoric acid and another type of molecule (choline, inositol, etc.) are attached. Wax esters are made of a fatty acid and a long chain alcohol and are a common form of lipid storage in certain species zooplankton . The main role of triglycerides is in the storage of lipids (fatty acids). Phospholipids are responsible for the structure of cell membranes (lipid bi-layer). Fatty acids are the main active components of dietary lipids. Fish are unable to synthesize fatty acids with unsaturation in the n-3 or n-6 positions yet these types of fatty acids are essential for many functions. These two types of fatty acids are, therefore, essential for the animal and must be supplied in the diet.
Deficiency in essential fatty acid result in general, in reduction of growth and a number of deficiency signs, including depigmentation, fin erosion, cardiac myopathy, fatty infiltration of liver, and "shock syndrome" (loss of consciousness for a few seconds following an acute stress). Salmonids require about 0.5 to 1% long chain polyunsaturated n-3 fatty acids (EPA (20:5 n-3) and DHA (22:6 n-3)) in their diet. This amount is easily covered by ingredients of marine origins, such as fish meal and fish oil, which are always present in significant amounts in salmonid feeds.
Carbohydrates represent a very large variety of molecules. The carbohydrate most commonly found in fish feed is starch, a polymer of glucose. Salmonid and many other fish have a poor ability to utilize carbohydrates. Raw starch in grain and other plant products is generally poorly digested by fish. Cooking of the starch during pelleting or extrusion, however, greatly improves its digestibility for fish. However, even if the starch is digestible, fish only appear to be able to utilize a small amount effectively. Carbohydrates only represent a minor source of energy for fish. A certain amount of starch or other carbohydrates (e.g. lactose, hemicellulose) is, nevertheless, required to achieved proper physical characteristic of the feed.
The vitamins are generally defined as dietary essential organic compounds, required only in minute amounts, and which play a catalytic role and but no major structural role. So far, 4 fat-soluble and 11 water-soluble vitamins or vitamin-like compounds have been shown to be essential to fish. Requirement is generally measured in young fast growing fish. However, requirements may depend on the intake of other nutrients, size of the fish, and environmental stress. The recommended levels and the deficiency signs are summarized in Tables 3 and 4. Many symptoms of vitamin deficiency are non-specific. It is also tedious and expensive to analyze diets for vitamins. Therefore, diagnostic of vitamin deficiencies is often difficult. Nutritional disorders caused by vitamin deficiencies can impair utilization of other nutrients, impair the health of fish, and finally lead to disease or deformities. Nutritional deficiencies signs usually develop gradually, not spontaneously. However, the culturist may obtain clues of deficiency indirectly through low feed intake and poor live weight and feed efficiency.
Table 3. Vitamin requirement of salmonids.
Table 4. Deficiency signs associated with various nutrients.
Inorganic elements (minerals) are required by fish for various functions in metabolism and osmoregulation. Fish obtain minerals from their diet but also from their environment. Many minerals are required in trace amounts and are present in sufficient quantity in the surrounding water for the fish to absorb through their gills. In freshwater, there is generally sufficient concentration of calcium, sodium, potassium and chloride for the fish to absorb and cover its requirements. The totality of the requirement for other minerals must, in general, be covered by the diet. Dietary minerals play many roles. There generally have a structural (e.g. bone formation) or catalytic (e.g. metalloenzyme) role. Minerals required by fish included calcium, phosphorus, sodium, potassium, magnesium, iron, copper, zinc, cobalt, selenium, iodine, and fluorine. The recommended levels of minerals in the diet are shown in Table 5. There are numerous deficiency signs and some are highlighted in Table 4. Reduced growth, feed efficiency and skeletal deformities is the most common signs of mineral deficiencies.
Table 5. Mineral requirement of salmonid fish in freshwater.
* Requirement in the absence of significant amounts of the specific mineral in the water.
Digestive system of fish is, in general, relatively simple compared to digestive system of birds and mammals but there are numerous similarities.
Barbel taste buds
Mouth teeth, no chewing, taste buds
Pharynx pharyngeal teeth (calcified structures)
Oesophagus short, thick, taste buds, gizzard
Stomach present or absent, acid, enzymes, rate of digestion correlates with mass of food remaining in stomach, emptying is affected by temperature.
Anterior intestine secretive and absorptive epithelial cells, no villi but numerous folds present, microvilli present, enterocytes with brush border membrane
Pyloric caeca present in some case, variable # between species and individuals (rainbow trout 50-200 p.c.). Increase absorptive surface, number apparently shows weak correlation with digestibility and growth. Ratio intestine/fork length = 0.7, ratio intestine + p.c./fork length = 3.9
Pancreas Generally a diffuse tissue, except for eel, pike, flat fish
Hindgut Not really morphologically distinct from anterior intestine but cell type changes. Squamous epithelial cells, mucus production, highly vacuolated cells, absorption of macromolecules by pinocytosis (tissue reabsorb proteins (enzymes) for recycling).
Table 6 presents the apparent digestibility coefficients for commonly used ingredients in salmonid feeds as measured by Cho et al. (1982). Fish have different digestive capabilities compared to terrestrial animals, and many feedstuffs, particularly cereal grains and their by-products which contain high levels of starch and fiber, are very poorly digested by carnivorous fish. The apparent digestibility of good quality protein by fish is very high. However, several factors can affect the digestibility of protein. The type of drying technique used during processing is a very important factors. A good demonstration of this is seen in blood meal. The protein digestibility of flame-dried blood meal is very low whereas the digestibility of spray-dried blood meal is very high. The same phenomenon can occur with fish meal.
Table 6. Apparent digestibility coefficients of ingredients measured with rainbow trout.
Diet formulation and preparation are the process of combining feed ingredients to form a mixture that will meet the specific goals of production. It is often a compromise between the ideal formula and practical considerations. The primary objectives are to produce a mixture that (is) :
balanced (to support maintenance, growth, reproduction, health)
Practical considerations :
Table 7. Composition of the grower formulae used by the OMNR Fish Culture Stations over the past 10 years.
The first consideration for formulation and production of successful diets is the quality of the feed ingredients. Diets produced with poor quality raw materials and under adverse processing conditions have inferior nutritive value and adverse effects on fish health. Quality criteria for the ingredients must be respected to insure that the final product is of consistent quality and that deleterious effects are avoided. The chemical composition (nutrient, energy, antinutrients, contaminants) of the ingredient obviously plays a determinant role the quality. However, biological aspects, such as digestibility and utilization of nutrients are most important and often overlooked.
The loss of indigestible matter from the diet as feces is the primary reason for variation in the nutritional value of feed ingredients. Measurement of digestibility provides, in general, a good indication of the availability of energy and nutrients, thus providing a rational basis upon which diets can be formulated to meet specific standards of available nutrient levels. Several factors can affect the digestibility of protein or specific amino acids. The type of drying techniques used during processing, the composition of the protein fraction are the factors which have a determinant effect on the digestibility of protein of a feed ingredients.
There are various qualities of fish meals on the market, relating to the original raw fish quality, level of ash in the meals, and the type of processing techniques used. The most important factor is the freshness of the product. Fish must be processed as soon as possible after capture. Ageing and spoilage decrease the nutritive value and also lead to the contamination with potential toxic compounds, such as histamine, cadaverine, and agmatine. The second most important factor is the type of raw material used (whole fish or by-products). By-products, such as those generated by the filleting industry (sometime referred to as white fish meal) have higher level of ash and lower level of protein than whole fish meals. High level of ash generally affects digestibility of dry matter and results in high waste outputs, and can also produce mineral imbalances (e.g. Zn deficiency).
The type of fish used is not necessarily a determinant factor in the quality of the products. At equal freshness and if the same processing technique is used, whole capelin, anchovy, herring, menhaden meals will support similar growth. During processing, the drying treatment is a key factor. Flame-dried products are less digestible and produce lower performances.
Table 8. Quality Standards of Fish Meal Required for Salmonid Diets.
Animal protein by-products can very useful complementary protein sources in fish diets. It is important to use highly digestible products with limited ash content. High ash content ingredients are generally more polluting and the ash dilute useful nutrient. It is especially important when buying these products to deal with suppliers who consistently provide high quality products. Apparent digestibility of animal by-product is relatively high (Table 6) and they have been used at significant levels in practical diet with success. For blood meal, the type of drying is of primary importance. Spray-drying produce the best results.
There are several plant proteins and grain by-products that are used on a regular basis in fish diet formula. Certain plant protein products have a good nutritional value (high in digestible protein, good amino acid profile) and are economical at the same time. Other products improve the physical characteristics of the pellets. The incorporation of certain products must be limited for various reasons, such as their content in starch and fibre, the presence of antinutritional or undesirable factors and their acceptability (palatability).
Many plant products contain antinutritional factors. Most plant protein ingredients are heat treated during processing, which greatly reduce the level of several antinutritional factors, such as soybean trypsin inhibitors. Excess heat, however, generally decreases the nutritional quality of plant protein products by destroying amino acids.
Fish diets formulated with high levels of certain plant protein ingredients appear to be nutritionally adequate but not very acceptable to certain fish species. For example, diets containing high levels of soybean meal are poorly accepted by chinook salmon and other salmonids. Recent experimental evidences suggest that soyasaponins may be a factor affecting performance of salmonids fed soybean meal.
Corn gluten meal is a plant protein ingredients known to be highly palatable for salmonids. Studies with rainbow trout and Atlantic salmon show that it complements soybean meal very well nutritionally. Recent results from our laboratory showed that corn gluten meal or combination of corn gluten meal and soybean meal can replace most of the fish meal without any effect on performance of the fish. Nonetheless, the incorporation of corn gluten meal must be limited in food fish production feeds due to its high concentration in xanthophylls which can produce undesirable pigmentation of the skin and flesh and may compete with expensive synthetic pigment added in the feed. However, recent evidences from our laboratory do not support this hypothesis.
Fish oil is the main source of lipid in salmonid diet. Marine fish oils are, in general, excellent sources of long chain n-3 PUFA (EPA & DHA), fatty acids required by salmonid. Other types of oils and fats can be used in salmonid diets. Vegetable (canola, soya, safflower, etc.) oils and animal fats (tallow, lard, poultry fat) can also be used at certain levels in feeds without effect on growth performance and health of the fish.
7.6.1 Rancidity problems:
Marine oils are rich in polyunsaturated fatty acids and are susceptible to rancidity. In all circumstances rancid oil must be avoided in the preparation of fish feeds. Rancid fat has deleterious effect on some of the nutrients present in fish feed and health of the fish. fatty liver disease is usually seen in fish fed rancid fat. Histologically, the main feature is the extreme infiltration of hepatocytes by lipids. Peroxide (PV), thiobarbituric acid (TBA) and anisidine (AV) values are in general parameters used to determine the degree of rancidity of lipid sources. Acceptable quality parameters for fish oil as suggested by Cho et al. (1983) are presented in Table 9. There is no unequivocal technique to measure rancidity and there is still doubts about the reliability of PV, AV and TBA value. High PV, AV or TBA suggest problems of lipid deterioration but are not always indicative of harmful rancidity. The easiest way to determine if a feed is rancid may be its smell. Feed with a rancid smell must not be fed. It is preferable to discard such feeds instead of jeopardising the health of the fish by feeding them.
Table 9. Quality Standards of Oils and Lipid in Final Product Required for Salmonid Diets
The are several forms of fish feed, including wet, moist, and steam-pelleted and extruded dry pellets. However, two basic types of formulated feed are generally used in intensive fish culture: dry and semi-moist diets. The diets are similar, the basic difference being that semi-moist pellets contain a larger proportion of raw fish and by-products which contribute a higher moisture level to the final product. Moist feeds have some merit in coastal regions where fresh raw fish and by-products are regularly available and economical. It is also possible that the physical characteristics of moist pellets are more palatable to some fish species. However, there is no evidence that such feeds are nutritionally superior to dry feeds. Moist feed may contain pathogens since the feed ingredients are only submitted to moderate heat treatment (pasteurization). In contrast to moist diets, dry feed are heat-treated and generally free from pathogens. They are also easier to transport and store. The bulk purchase and storage of quality dry ingredients is possible and ensures a continuous supply of quality feed. The dry ingredients on the commodity market are more quality defined than raw fisheries products and can be supplied regularly. Hence it is possible to formulate dry feeds more precisely with the available knowledge of fish nutrition. Most nutrient in dry feeds are stable are room temperature and therefore dry feeds can be stored safely without freezing for periods which depend on storage conditions (approx. 3 months in a cool, shady, and well-ventilated location).
Widely used dry feeds today may divide into three types: (1) steam-pelleted feed; (2) partially extruded, slow-sinking pellets, and (3) expanded and floating pellets. Feeding dry pellets either by hand or with automatic feeders is much simpler than that of moist feeds. The problem of acceptability of dry feeds by some fish species can usually be solved by better feeding techniques and fish culture management. Otherwise, fry which have difficulty in accepting dry feeds can be started with semi-moist feed and gradually shifted over to dry feed within 3-5 weeks.
A formulated dry fish feed must be pelleted and/or crumbled so as to be durable and water stable. Formulated feeds must also have desirable physical and textural characteristics, and be of the correct sizes to be readily acceptable by different sizes of fish. Disintegrated and uneaten feed pollutes the water and creates stresses from low oxygen and high nitrogen and organic wastes, with serious effects on growth and health. Some of the important factors in manufacturing a durable, dry fish feed without fines are (1) physical properties of the ingredients, (2) particle size of ingredients, (3) conditioning time and temperature in the pellet mill, (4) quality of steam supply, (5) compression pressure through the die, and (6) efficiency of sifting/grading and fat-spraying equipment. Many of the dietary problems experienced in fish culture in the past have been related to the physical quality of the pellets and granules, which was in turn related to poor quality ingredients, inadequate manufacturing processes, and negligent practices. Unfortunately for fish feed, the manufacturing process is of crucial importance. Having to transfer dietary nutrients into the fish through the water medium presents problems which are unknown in other animal-feeding practices. Therefore, all newly opened bags should be checked for the presence of excess fines, undersized granules, durability, foreign particles, too little or too much oil, mildew, and other evidence of poor quality. Any bag or batch of feed judged to be questionable and any with a detectable "rancid" smell should not be fed. All questionable feed should be immediately reported to a qualified nutritionist and returned to the manufacturer for replacement.
Table 10. Recommended particle size for salmonid diets
Feeding systems may be defined as all feeding standards and practices employed to deliver nutritionally balanced and adequate amount of diets to animals, so maintaining normal health and reproduction together with efficient growth and/or work performance. Until now the feeding of fish has been based mostly on folkloric practices while the main preoccupation has been to develop "magic" diet formulae. Many "hypes" such as mega-fish meal and mega-vitamin C diets have come and gone, and we are now in the age of the "Norwegian Fish Doughnut" (>36% fat diet)! Whichever diet one decides to feed, the amount fed to achieve optimum or maximum gain is the ultimate measure of one’s productivity in terms of biological gain, economical benefit and/or environmental sustainability.
Scientific approaches have been used in the feeding of land animals for over a century. The first feeding standard for farm animals was proposed by Grouven in 1859, and included the total quantities of protein, carbohydrate and ether extract (fat) found in feeds, as determined by chemical analysis. In 1864, E. Wolf published the first feeding standard based on the digestible nutrients in feeds.
Empirical feeding charts for salmonids at different water temperatures were published by Deuel and his colleagues and were likely intended for use with meat-meal mixture diets widely in use at that time. Since then several methods of estimating daily feed allowance have been reported. Unfortunately all methods have been based on the body length increase or live weight gain, and dry weight of feed and feed conversion, rather than on biologically available energy and nutrient contents in feed in relation with protein and energy retention in the body. These methods are no longer suitable for today’s energy- and nutrient-dense diets, especially in the light of the large amount of information available on the energy metabolism of salmonids.
Many problems are encountered when feeding fish, much more so than with feeding domestic animals. First, delivery of feed to fish in a water medium requires particular physical properties of feed together with special feeding techniques. It is not possible in the literal sense to feed fish on an "ad libitum" basis, like it is done with most farm animals. The nearest alternative is to feed to "near-satiety" with very careful observation over a pre-determined number of feedings per day; however, this can be very difficult and subjective. Feeding fish continues to be an "art" and the fish culturist, not the fish, determines "satiety" as well as when and how often fish are fed. The amount of feed not consumed by the fish can not be recovered and, therefore, feed given to them must be assumed eaten for inventory and feed efficiency calculations. This can cause appreciable errors in feed evaluation as well as in productivity and waste output calculations. Meal-feeding the fish pre-allocated amounts by hand or mechanical device based on theoretical energy requirement may be the only logical choice. Uneaten feed represents an economical loss and becomes 100% solid and suspended wastes! Meal-feeding a pre-allocated amount of feed calculated based on the theoretical energy requirement of the animal may not represent a restricted feeding regime as suggested by some since the amount of feed calculated is based on the amount of energy required by the animal to express its full growth potential.
There are few scientific studies, based on nutrition and husbandry, on feeding standards and practices; however, there are many duplications and "desktop" modifications of old feeding charts with little or no experimental basis. Since the mid-1980's, development of high fat diets has led to most rations being very energy-dense, but feeding charts have changed little to reflect these changes in diet composition. Most feeding charts available today tend to over-estimate feed requirements and this overfeeding has led to poor feed efficiencies under most husbandry conditions, and this represent a significant, yet avoidable, waste of resources for aquaculture operations. In addition, it may results in self-pollution which in turn may affect the sustainability of aquaculture operations. Recent governmental regulations imposing feed quota, feed efficiency guidelines and/or stringent waste output limit may somewhat ease the problem. Sophisticated feed management systems, such as underwater video camera or feed trapping devices, have been developed to determine fish satiation or the extent of feed wastage and are promoted by many as a solution to overfeeding. However, regardless of the feeding system or method used, accurate growth and feed requirement models are needed in order to forecast growth and objectively determine biologically achievable feed efficiency (based on feed composition, fish growth, composition of the growth). These estimates can be used as yardsticks to adjust feeding practices or equipment and to compare results obtained.
The development of scientific feeding systems is one of the most important and urgent subjects of fish nutrition and husbandry because, without this development, nutrient dense and expensive feeds are partially wasted. Sufficient data on nutritional energetics are now available to allow reasonably accurate feeding standards to be computed for different aquaculture conditions. Presented here is a summarized review of the basis of a nutritional energetic approach to estimating feed requirement and waste output of fish culture operation as well as the development of the Fish-PrFEQ computer program. Results obtained from a field station are presented and provide a framework to examine the type of information that can be derived from bioenergetic models and generate a feed requirement scenario for the next production year.
Evaluating and/or predicting growth performance of a fish culture operation or a stock of fish firstly requires production records of past performance. These records may become databases for calculating growth coefficients, temperature profiles during growth period and feed intake and efficiency for various seasons etc. One such production records for a lot of rainbow trout from a field station is shown in Table 11. A lot of 100 000 fish was reared over a 14-month (410 days) production cycle between May, 1995 and June, 1996. Cumulated live weight gain (fish production) was 72 tonnes with feed consumption of 60 tonnes which gave an overall feed efficiency (gain/feed) of 1.19 (ranged between 1.11 – 1.22). Water temperature ranged from 0.5°C in winter to 21°C in summer which is typical of most lakes in Ontario. In spite of the wide fluctuation in water temperature, the thermal-unit growth coefficients (TGC) was fairly stable ranging between 0.177 – 0.204. Total mortality was around 9% over 410 days.
From the production record (Table 11) one can extrapolates an overall growth coefficient of 0.191 and this coefficient can be used for the growth prediction of next production cycle with assumption of similar husbandry conditions and fish stock are used. Total feed requirement and weekly or monthly feeding standards can be computed on the basis of this growth predictions plus the quality of feed purchased.
Table 11. - Fish production records from a field station
Using production records as a starting point, feed requirements and waste output can scientifically be estimated based on the following three concepts:
1) Prediction of
growth and nutrient and energy gains
Accurate prediction of growth potential of a fish stock under given husbandry condition is an inevitable prerequisite to the estimation of energy or feed requirement (e.g. weekly ration). The formula most commonly used for fish growth rate expression is instantaneous growth rate known as "specific growth rate (SGR)" which is based on the natural logarithm of body weight:
SGR = (ln FBW - ln IBW) / D. (1)
FBW is final body
SGR has been widely used by most biologists to describe growth rate of fish. However, the exponent of natural logarithm underestimates the weight gain between the IBW and the FBW used in the calculation and it also grossly overestimates predicted body weight at weights greater than FBW used. Furthermore the SGR is dependent on the IBW, making meaningless comparisons of growth rates among different groups unless IBW are similar.
A more accurate and useful coefficient for fish growth prediction in relation to water temperature is based on the exponent 1/3 power of body weight. Such a cubic coefficient has been applied both to mammals and to fish. The following modified formulae were applied to many nutritional experiments:
T is water temperature (° C)
(NOTE: 1/3 exponent must contain at least 4 decimals (e.g. 0.3333) to maintain good accuracy)
This model equation has been shown by experiments in our laboratory and several field stations to represent very faithfully the actual growth curves of rainbow trout, lake trout, brown trout, chinook salmon and Atlantic salmon over a wide range of temperatures. Extensive test data were also presented by Iwama and Tautz (1981). An example of the relationship among growth, water temperature and TGC is shown in Figure 11. Growth of some salmonid stocks used for our experiments in freshwater gave the following TGC:
Since these TGC values and growth rate are dependent on species, stock (genetics), nutrition, environment, husbandry and others factors, it is essential to calculate the TGC for a given aquaculture condition using past growth records or records obtained from similar stocks and husbandry conditions.
Once the expected TGC and water temperature profile during the production period are established, expected live weight gain (LWG) and recovered energy (RE), nitrogen (RN) and phosphorus (RP) on basis of dry matter (DM, 20-35% of live body weight) in carcass can be computed in the following manners:
LWG = FBW - IBW (4)
RE (or RN, RP) = LWG x DM x GE (or N, P) (5)
LWG is live weight
Because of a large proportion of the nutrients (e.g. protein, lipid) and, consequently of the dietary energy, consumed by fish is retained as carcass body constituents, carcass energy gain is a major factor driving dietary energy requirement of the fish. Carcass moisture, protein and fat contents in various life stages dictate energy level of fish. These factors are influenced by species, genetics, size, age and nutritional status. The dry matter and fat contents of the fish produced are, in general, the most variable factors and have a determinant effect on energy content of the fish. For example, relatively fatty Atlantic salmon and rainbow trout may require more dietary energy per unit of live body weight than leaner salmonids such as brown trout, lake trout and charr. Fish containing less moisture (more dry matter) and more fat require more energy allocation in feeding standards.
The simplistic assumption of the constant body composition within a growth stanza in certain published models is not necessarily valid for different species and sizes. Dry matter and energy content of fish can increase dramatically within a growth stanza, especially in the case of small fish. Underestimation or overestimation of the feed requirement is likely to occur if constant carcass energy content is assumed in calculations. Reliable measurements of carcass composition of fish at various size are essential. Nutrient and energy gains should be calculated at relatively short size intervals as possible, at least for small fish (<100 g). Additionally, composition of the diet, notably the digestible protein to digestible energy ratio and the lipid content of the diet, can have a very significant influence on the composition and energy content of the carcass. Estimation of carcass composition and energy content should rely on data obtained with fish fed diets similar to those one intends to use.
Waste output loading from aquaculture operations can be estimated using simple principles of nutrition and bioenergetics. Ingested feedstuffs must be digested prior to utilization by the fish and the digested protein, lipid and carbohydrate are the potentially available energy and nutrients for maintenance, growth and reproduction of the animal. The remainder of the feed (undigested) is excreted in the feces as solid waste (SW), and the by-products of metabolism (ammonia, urea, phosphate, carbon dioxide, etc.) are excreted as dissolved waste (DW) mostly by the gills and kidneys.
The total aquaculture wastes (TW) associated with feeding and production is made up of SW and DW, together with apparent feed waste (AFW):
TW = SW + DW + AFW (6)
SW, DW and AFW outputs are biologically estimated by:
SW = [Feed consumed x (1-ADC)] (7)
DW = (Feed consumed x ADC) - Fish produced (nutrients retained) (8)
AFW = Actual feed input – Theoretical feed requirement (9)
in which ADC is the apparent digestibility coefficients of diets. Measurements of ADC and feed intake provide the amount of SW (settled and suspended, AFW-free) and these values are most critical for accurate quantification of aquaculture waste. ADC for dry matter, nitrogen and phosphorus should be determined using reliable methods by research laboratories where special facility, equipment and expertise are available. More information on the equipment and procedures may be obtained from the website www.uoguelph.ca/fishnutrition.
DW (N or P) can be calculated by difference between digestible N or P intake and retained N (RN) or P (RP) in the carcass if this information is available, or by using a digested nutrient retention efficiency (NRE = Retained/Intake). Reliable NRE are necessary and should be determined or estimated for each type of diet used by research laboratories where expertise is available. However, controlled feeding and growth trial(s) with particular diets at production sites are essential to validate and fine-tune the coefficients from the laboratory. Dissolved nitrogen output depends very much on dietary protein and energy ratio and amino acid balances and rate of protein deposition by the fish, therefore all coefficients must be determined on a regular basis, particularly when feed formulae are changed. Assuming constancy of many coefficients is a dangerous exercise.
Accurate estimation of total solid waste (TSW) requires a reliable estimate of AFW. Feeding the fish to appetite or near satiety is very subjective and unfortunately TW contains a considerable amount of AFW under most fish farming operations. The use of "biomass gain x feed conversion" as an estimate of real feed intake of the fish to calculate waste output used in certain waste output prediction models can grossly overestimate the feed intake in many operation where overfeeding is common and result in an underestimation of the TSW output.
It is very difficult scientifically to determine the actual feed intake by fish in spite of many attempts (mechanical, radiological and biological) that have been made by biologists. Since estimation of AFW is difficult and almost impossible, the best estimates can be made based on energy requirements and expected gain in which the energy efficiency (energy gain/intake) indicates the degree of AFW for a given operation. The theoretical feed requirement (TFR) can be calculated based on nutritional energetic balance as follows:
TFR = Retained + Excreted (10)
and the amount of feed input above the TFR should be assumed as AFW and all nutrient contents of the AFW must be included in solid waste quantification. This approach may yield relatively conservative estimates.
Biological procedures based on the ADC for SW and comparative carcass analyses for DW were shown to provide very reliable estimates. Biological methods are flexible and capable of adaptation to a variety of conditions and rearing environments. It also allows estimation of the theoretical feed requirement and waste output under circumstances where it would be very difficult or impossible to do so with a chemical/limnological method (e.g. cage culture). Properly conducted biological and nutritional approaches to estimate aquaculture waste outputs are not only more accurate but also more economical than chemical/limnological method.
The waste outputs from the field station (see Table 11) are tabulated in Table 12. SW was estimated at 10 610 kg (fish production 72 t; 60 t feed input over 14 months). SW represented 90% of TSW, since AFW (actual feed input – theoretical feed requirement) was estimated at 1 201 kg or 2 % of feed input (60 277 kg in Table 11). The TSW outputs were equivalent to 164 kg per tonne fish produced. Phosphorus waste was 5.11 kg/t fish produced and nitrogen 30.64 kg. Total water consumption during 14 months was 13 469 m3, therefore the average effluent quality can be estimated at: solid 0.877 mg/L, nitrogen 0.163 and phosphorus 0.027 (Table 12). The diet (MNR-91HG) and the procedures to estimate waste production as well as comparative data of chemical and biological estimations from field experiments at the Ontario Ministry of Natural Resources (OMNR) Fish Culture Stations are described elsewhere (Cho et al., 1991, 1994).
Table 12 - Waste outputs and effluent quality from fish production operation in Table 1
* Actual amount of feed fed – Theoretical amount of feed required
A relatively large portion of dietary energy is expended for maintenance or basal metabolism, which is the minimum energy and nutrients required necessary to maintain basic life processes. Maintenance energy requirement is approximately equal to the heat production of a fasting animal. This amount of dietary energy represent as an absolute minimum of "energy-yielding" nutrients must be covered before any nutrients can be used for growth and reproduction of the animal. Otherwise body tissues will be catabolized because of a negative energy balance between intake of dietary fuels and energy expenditure. \
A review of available data suggest that a HEf of about 36-40 kJ/kg0.824 per day appear accurate for rainbow trout at 15°C, at least for fish between 20 and 150 g live weight with which most of studies have been conducted. Water temperature has a major influence on basal metabolism of fish. The following equation to estimate HEf of salmonids as a function of water temperature (10):
HEf = (- 0.0104 + 3.26T - 0.05T2) (BW 0.824 ) D -1 (11)
fasting heat production in kJ
Ingestion of food by an animal which has been fasting results in an increase in the animal's heat production, this heat production is known as heat increment of feeding (HiE). The physiological basis of this increased heat production includes the post-absorptive processes related to ingested food, particularly protein-rich food and the metabolic work required for the formation of excretory nitrogen products, as well as the synthesis of proteins and fats in the tissues from the newly absorbed, food-derived substrates such as amino acids and fatty acids.
The HiE of rainbow trout fed a balanced diet was observed to be approximately 30 kJ/g digestible N or the equivalent of 60% HEf (Cho and Kaushik, 1990), but these relationships do not always hold true. Studies with farm animals suggest that HiE associated with growth may be more appropriately quantified as a factorial function of protein and lipid deposition rates. Protein and lipid oxidation rates also appear to contribute to HiE (Cho et al., 1982). Experimental observations suggest that HiE is approximately equivalent of 17% of net energy intake, i.e. 0.17(RE+HEf) for rainbow trout and other salmonids. This value is used in the bioenergetic model presented here. Studies are underway to quantify HiE as a function of protein and lipid deposition and oxidation rates.
Biological oxygen requirement of feeding fish is equal to the total heat production (HEf + HiE / Qox) in which the oxycalorific coefficient (Qox) used in the model is 13.64 kJ energy per g oxygen. This represent the absolute minimum quantity of oxygen that must be supplied to the fish by the aquatic system. Oxygen requirement per unit of BW per hour will vary significantly for different fish sizes, water temperatures and growth rates.
The calculation of total energy requirement and consequently feed allocation of the animal can be accomplished as follows:
1. Calculation of expected live weight gain (LWG = FBW - IBW) and recovered energy (RE) based on carcass dry matter content (DM = 20-35% of live body weight and gross energy (GE) contents = 25-30 kJ/g DM):
RE = LWG x DM x GE (12)
2. Allocation of approximate maintenance or fasting energy requirement at a given water temperature (T):
HEf = (- 0.0104 + 3.26T - 0.05T2) (BW 0.824) d-1 (11)
3. Allocation of approximate heat increment of feeding for maintenance and growth ration:
HiE = (RE + HEf) x 0.17 (13)
4. Allocation of approximate non-fecal energy loss:
ZE + UE = (RE + HEf + HiE) x 0.09 (14)
5. Theoretical/minimum energy requirement:
TER = 1) + 2) + 3) + 4) = [(RE + HEf) x 1.2753] (15)
6. Feed allowance or feeding standard:
FA = TER / DE x Qfi (16)
theoretical/minimum energy requirement (MJ)
Qfi is an adjustment factor determined by the fish culturist to provide flexibility for estimating realistic FA under a given husbandry condition (if one observes that more or less feed may be required than predicted by the model). The minimum digestible energy requirement that should be fed to the fish is the sum of retained energy (RE) and energy lost as HEf + HiE + ZE + UE. The amount of feed can be estimated on a weekly or monthly basis, and recalculated if any parameter (growth rate, water temperature, etc.) is changed. The computed quantity of feed should be regarded as a minimum requirement under most conditions and fish culturists should fine-tune the feeding level to own local conditions using the adjustment factor (Qfi).
The overall energy cost of producing one kg of rainbow trout is around 15-16 MJ DE, but this ranges from 10 MJ for fry to more than 20 MJ for fish of near 3 kg. Even though maintenance energy requirement per kg BW is much higher in small than in large fish, overall energy cost of production is much higher in large fish because of high "growth-fattening cost". This may become much more significant when feeding overly high energy (fat) diets, hence more than 50 kJ DE per g DP (or less than 20 g DP/MJ DE) is not recommended. Water temperature greatly affects heat production and oxygen consumption of poikilotherms and a growing fish of 100 g is expected to consumed 110, 210 and 300 mg oxygen/kg BW/hr at 5, 10, 15°C, respectively.
Table 13 summarizes the monthly fish sizes and feed rations predicted by the bioenergetic models program for the field station based on the production records (see Table 11). The feed requirements were calculated using a single TGC (0.191) for the whole production cycle (14 months) and actual temperature profile. The nutrient and energy gains used in the calculations were based on carcass composition values for rainbow trout of various sizes obtained in different laboratory trials at the University of Guelph. Nutrient and energy retention efficiencies (NRE and ERE) used were derived from previous studies at another fish culture station (Harwood Fish Culture Station, Harwood, Ontario) using comparable diets (Cho et al., 1994). The main discrepancy is between the actual and predicted feed amount for the first four months with actual feed input being greater than predicted allocation. This may indicate that overfeeding occurred in 1995, however, real feed intake by the fish could be somewhere between the predicted amount and the actual amount. Using this information, the fish culturist can fine-tune the program in the next production cycle. In the remaining 10 month, the feed allocation estimated by the model was very close to the actual feed fed, the largest discrepancies (in terms of predicted/actual) occurring at very low water temperature (0.5° C).
This simulation may not be considered a perfect example of independent or objective validation of the model but is, nevertheless, an adequate demonstration of the realism of the predictions from bioenergetic models. Most of the parameters used in the calculations are fairly independent from the actual data. For example, the carcass composition data were from a number of laboratory trials which had nothing to do with actual data. The TGC and the temperature profile used in the calculation are not independent from the actual data because it is essential to use actual values or values from previous production cycle if these are available and repeatable. TGC and water temperature are main inputs required from the fish culturist by the models. The predicted values from Table 11 were calculated a posteriori and their main use is as production scenario for following year based on 1995 production performance. The predicted values can also be used as yardsticks to compare the results obtained with what was predicted to be biologically achievable and adjust feeding practices or equipment in the following production cycle.
Table 13. - Prediction of fish body weight and feed requirement based on 1995 production records in Table 11.
** Overall TGC = 0.191 from Table 1 was used to p
** Overall TGC = 0.191 from Table 1 was used to predict body weight and total feed requirement
A stand-alone multimedia program for the Windows 95 ™ platform was developed in visual basic language with database functionality by the Ontario Ministry of Natural Resources. The program has 4 modules for fish production/growth prediction, waste output quantification, feed allowance estimation and oxygen requirement table and is based on the bioenergetic models presented above. Feed composition, body weight, water temperature, flow rate and mortality are entered by the user but waste, retention and other coefficients are parameters that are locked and may only be revised with an authorized program update diskette. These coefficients should be determined by qualified nutritionists from feed manufacturers or research institutions since specific coefficients are required for each type of diets. The use of unrelated coefficients result in under or overestimation of feed requirements and waste outputs. Live weight gain, feed efficiency, growth coefficients, solid, nitrogen, phosphorus in the effluent, total waste load, feed ration and oxygen requirements are some of the output parameters generated by the models.
Feed costs represent a very significant proportion of the production cost in salmonid fish culture. Many fish culture operations have poor feed efficiencies (gain/feed) and this contributes to the high cost of production and often results in significant water pollution. It is necessary to optimize feeding regimes to improve the economical and environmental sustainability of aquaculture.
It is not always clear if low feed efficiencies observed under certain conditions are due to feed wastage or due to a real decrease in feed utilization efficiency of the fish. The effect of feeding level on the efficiency of feed utilization in rainbow trout and other salmonids is the subject of controversy. It has been suggested that optimum feed efficiency is achieved at feeding levels below that required for maximum growth in salmonids. Other studies suggest that feed efficiency improves to its maximum at moderate feed restriction (e.g. 50% of maximum ration) and this optimum is maintained up to the ration required for maximum growth of the fish. It has also been suggested that maximum feed efficiency of fish is attained at maximum feed intake and maximum growth. Most of these observations are derived from studies using fixed ration (% live body weight) which may not represent the fish’s actual feed requirement or studies conducted under poorly controlled experimental conditions (mechanical distribution of feed, variable temperature, etc.).
While important from a production point of view, feed efficiency can be a misleading expression of nutrient and energy utilization. Physical quantity of feed used is not a measure of biologically available nutrients and energy supplied to the animal. In addition, weight gain does not always accurately reflect protein, lipid and energy gains since the composition of weight gain is often variable. Protein deposition is associated with substantial water deposition whereas lipid depots contain little water. The ratio between protein and lipid deposition will have an impact on live weight gain and, consequently, feed efficiency.
Being poikilothermic animals, the metabolic rate, growth, energy expenditure, and feed intake of fish are highly influenced by water temperature. It is, therefore, important to study how water temperature affects these parameters, as well as to determine the effect of temperature on the efficiency of nutrient and energy utilization. Studies have suggested that temperature can affect the efficiency of energy utilization in salmonids.
The effect of feeding level and water temperature on feed utilization was recently re-examined under highly controlled conditions (careful hand-feeding to avoid feed waste, controlled temperature, etc.). The results from the study indicated that fish consuming more feed as a result of an increase in water temperature or an increase feed allocation grew faster but appeared to utilise digestible nutrients with similar efficiencies (Table 14, Figure 2). Feeding level or water temperature had very little effect on feed efficiency. The main factors affecting feed efficiency of fish fed a balanced practical diet under practical is, therefore, feed wastage.
Feeding frequency and timing is another factor that has been suggested as affecting feed intake and utilization by fish. On a weekly basis, studies have suggested that feeding the equivalent of six days a week resulted in growth performance similar to feeding 7 days a week. Feeding five days a week resulted, however, in significantly less growth (Table 15). There is no good evidence that daily feeding frequency and timing affect feed utilization. The most important factor is to insure frequent and spaced enough meals to insure that the animal can consumed enough feed to meet its growth potential. This generally means more frequent feeding for fish of smaller size. Table 10 provides informal guidelines for daily feeding frequency (# of meal/day) as a function of fish size. There is, in general, slight "between" and "within" day variations in the appetite of fish, especially if they are free to choose when to feed (i.e. with a demand feeder). Fish will, however, easily adapt to a feeding schedule. Being attentive to changes in appetite of fish is, nevertheless, a very important skill fish culturist must acquire.
Table 14. Growth performance and feed efficiency of rainbow trout (IBW=13.3 g/fish) fed the experimental diet for 12 weeks at 3 feeding levels and 4 water temperatures. N = 3 for each feeding level within temperature and for each temperature.
NS = near satiation, R1, R2 = restricted diets. SEM = standard error of mean. HSD = Tukey’s honestly significant difference (P < 0.05). Means in the same column (within each temperature) with different superscripts (a, b, c) are statistically different (P < 0.05). The superscripts w, x, y, z are used for comparing results between temperatures at the NS feeding level (P < 0.05). 1weight as fed basis, 2FE = wet weight gain / dry feed
Table 15. Effect of weekly feeding frequency on growth and feed efficiency of rainbow trout.
Initial weight =