The study of the balance among dietary energy supply, expenditure, and gain offers a relatively simple way of looking at dietary component utilization by animals. Study of the energy transactions in animals requires that com- ponents be expressed in compatible terms. Classically, all measurements of energy transactions made by animal nutritionists were expressed in terms of calories. The calorie used in nutrition is the 15◦C calorie (the energy re- quired to raise the temperature of 1 g water from 14.5 to 15.5◦C). However, the joule (J) was adopted in the Syst`eme International des Unit´es (Inter- national System of Units) as the preferred unit for expression of electrical, mechanical, and chemical energy and by most nutrition journals as the basic unit for expressing dietary energy. One joule is defined as 1 kg-m2/sec2or 107erg. One 15◦C calorie is equivalent to 4.184 J.
∗ Editors note. The authors prefer to use the joule to measure energy content and reactions, whereas many other authors use the calorie for energy measurements. These are convertible:
1Cal=4.184 J, or 1 kcal=4.184 kJ. See below.
Many terms have been invented and applied to describe energy trans- actions occurring in animals. Historical terms, such as “specific dynamic action of food,” are still used, even though they imply nothing about the un- derlying relationships; others such as “work of digestion” have specific but incorrect implications regarding underlying relationships (Baldwin and By- water, 1984). Different groups have tended to adopt and defend alternative systems of nomenclature to describe the partition of energy in animals. This is especially apparent in fish biology, where nomenclatures and mode of expression of energy transaction are extremely diverse. In 1981, a subcom- mittee of the Committee on Animal Nutrition of the U.S. National Research Council was appointed to develop a systematic terminology for description of energy utilization by animals, including fish (NRC, 1981). This system is presented schematically in Fig. 1.1 and has been widely adopted by an- imal nutritionists. This rational nomenclature has also been adopted by a number of fish nutrition researchers and is used in this chapter. Its various components are discussed below.
1.4.1. Gross Energy: Dietary Fuels
Gross energy (GE) is the commonly used term for the enthalpy (H) of combustion in nutrition. However, as opposed to enthalpy, GE is generally represented by a plus (+) sign. The GE content of a substance is usually measured by its combustion in a heavily walled metal container (bomb) under an atmosphere of compressed oxygen. This method is referred to as bomb calorimetry. Under these conditions, the carbon and hydrogen are fully oxidized to carbon dioxide and water, as they arein vivo. However, the nitrogen is converted to oxides, which is not the casein vivo. The oxides of nitrogen interact with water to produce strong acids, an endergonic reac- tion. These acids can be estimated by titration, allowing a correction to be applied for the difference between combustion in an atmosphere of oxygen and catabolismin vivo(Blaxter, 1989).
The GE content of an ingredient or a compounded diet depends on its chemical composition. The mean GE values of carbohydrates, proteins, and lipids are 17.2, 23.6, and 39.5 kJ/g, respectively (Blaxter, 1989). Minerals (ash) have no GE because these components are not combustible. IE is the notation adopted by the NRC (1981) for an animal’s intake GE of (Fig. 1.1).
IE is simply the product of feed consumption and GE.
1.4.2. Fecal Energy and Digestible Energy
Before the feed components can serve as fuels for animals, they must be digested and absorbed (sometimes called “assimilated,” a term whose use
should be discouraged) from the digestive tract. Some feed components resist digestion, and these pass through the digestive tract to be voided as fecal material. Egestion (excretion through feces) of components contain- ing GE is referred to as fecal energy (FE) losses. The difference between the GE and the FE of a unit quantity of this diet is termed the digestible energy (DE). DEI was adopted by the NRC (1981) to represent the intake of DE, the product of feed intake and DE of the feed, or IE minus FE (Fig. 1.1).
Variation in the digestibility of foods is generally a major factor affecting the variation in their usefulness as energy sources to the animal, since FE is a major loss of ingested GE. Therefore, values for DE and values for the digestibility of individual nutrients should be used to estimate levels of available energy and nutrients (as opposed to GE or crude nutrients) in feed ingredients for diet formulation (Cho and Kaushik, 1990). Formulation on a GE or crude nutrients (e.g., crude protein) basis, rather than formulation on a DE or digestible nutrients basis, is still very common in fish nutrition, but sufficient information on DE values of common fish feed ingredients is now available to allow feeds to be formulated on a DE or a digestible nutrient basis. It is, however, important to emphasize that DE is only an indication of the potential contribution of the energy from nutrients in the ingredient.
These values do not serve as measures of the utilizable energy or of the productivity of the diet.
1.4.3. Measurement
The first task in the measurement of digestibility of feeds and feedstuffs is the collection of fecal samples. In aquatic animals, separating fecal material from water and avoiding contamination of the feces by uneaten feed neces- sitate the use of approaches that differ significantly from those commonly used to measure digestibility interrestial animals and birds.
Quantitative collection of fish feces is very difficult, and therefore, di- gestibility measurements using direct methods, involving total collection of fecal material, are rarely used with fish. Digestibility measurements in fish must, therefore, rely on the collection of a representative fecal sample (free of uneaten feed particles) and the use of a digestion indicator to obviate the need to quantify dietary intake and fecal output (indirect method). The inclusion of a digestion indicator in the diet allows the digestibility coeffi- cients of the nutrients in a diet to be calculated from measurements of the nutrient-to-indicator ratios in the diet and feces (Edin, 1918).
Several techniques have been used to collect fecal material from fish. The suitability of these various techniques has been a subject of discussion and disagreement among fish nutritionists for many years (Smith et al.,1980;
Cho et al.,1982; Cho and Kaushik, 1990; Hajenet al.,1993a; Smithet al.,
1995; Guillaume and Choubert, 1999). Some early, yet still widely used, techniques are the collection of feces from the lower part of the intestine by stripping (Nose, 1960), by suctioning fecal material, or by dissecting the fish (Windellet al.,1978). It is generally agreed that forced evacuation of fecal material from the rectum results in the contamination of the sam- ples with physiological fluids and intestinal epithelium that would otherwise have been reabsorbed by the fish before natural defecation. This affects the reliability of this type of approach and, in general, leads to underestimation of digestibility (Choet al.,1982; Hajenet al.,1993; Guillaume and Choubert, 1999).
Techniques involving the collection of feces voided naturally by the fish are, therefore, preferable. Smith (1971) developed a metabolic chamber to collect feces samples voided naturally into the water by fish. With this method, the fish need to be force-fed, and they frequently regurgitate and may not be in a positive nitrogen balance status. This technique clearly im- poses an unacceptable level of stress on the fish and produces estimates of digestibility of questionable reliability (Choet al.,1982). Other techniques, such as the periodical collection of feces by siphoning from the bottom of a tank, are also likely to yield inaccurate estimates of digestibility since the breakup of feces by fish movement may lead to leaching of nutrients and, therefore, overestimation of digestibility of nutrients.
To prevent these problems, specific devices were developed by Oginoet al.
(1973), Choet al. (1975), and Choubertet al. (1979) to collect fecal material passively. Ogino et al. (1973) collected feces by passing the effluent water from fish tanks through a filtration column (TUF column). Cho and Slinger (1979) developed a settling column to separate the feces from the effluent water (Guelph system) and Choubertet al.(1979) developed a mechanically rotating screen to filter out fecal material (St. P´ee system). These systems are convenient and have been adopted in many laboratories around the world. They are widely recognized as producing meaningful estimates of digestibility of nutrients if used correctly, despite the fact that differences of opinion about the accuracy of these systems remain. In a study comparing the TUF column and the Guelph system, very similar apparent digestibility coefficients (ADC) of dry matter, protein, lipid, and energy were obtained with both methods for two reference diets (Satohet al.,1992).
It is clear that differences exist in the estimates of digestibility with the various techniques currently used (Choet al., 1982). It is difficult to reach objective conclusions about the accuracy and reliability of the various tech- niques, as there are relatively few solid experimental studies allowing seri- ous comparisons. Direct measurements of energy and nutrient deposition and various losses (nonfecal losses, heat production, etc.) are virtually the
only way of objectively comparing the accuracy of the various approaches.
However, measurements of the various components of the energy or nu- trient budgets (e.g., nonfecal losses, heat production) of fish also require specific expertise and are subject to errors.
The differences in estimates of apparent digestibility measured with the most common techniques (stripping, St. P´ee system, TUF column, Guelph system) tend to be fairly stable when these techniques are used in a stan- dardized fashion. This suggestion comes from examination of the results of studies examining energy or nutrient depositions of groups of fish at differ- ent measured intakes of various practical diets (Kaushiket al.,1981; M´edale et al., 1995; Azevedoet al.,1998; Ohta and Watanabe, 1998; M´edale and Guillaume, 1999; Rodehutscord and Pfeffer, 1999). Regressions of energy and N depositions as a function of DE or digestible nitrogen (DN), measured with different techniques (stripping, St-P´ee system, TUF column, Guelph sys- tem), show very significant linear relationships within studies (R2 > 0.96).
This suggests that digestibility measurements appear to be consistent within techniques and that, if investigators adopt one technique and apply it in a standard fashion, very meaningful (informative) energy or nutrient budgets can be constructed.
1.4.4. Apparent versus True Digestibility
Feces are composed of the undigested food components and the unreab- sorbed residues of body origin. These residues are the remains of mucosal cells, digestive enzymes, mucoproteins, and other secretions released into the digestive tract by the animal, together with the residues of the microflora which inhabit the digestive tract (Nyachoti et al.,1997). The enthalpy of combustion of these materials represents a loss of energy which is not de- rived from the food. This energy loss is designated fecal energy of metabolic origin (FmE) and is influenced by the characteristics of the food and the level of feed intake. Estimates of FmE allow the description of “true” di- gestible energy values, which are greater than “apparent” digestible energy values. The term “true” digestibility may be misleading since, to the animal, FmE losses are real and inevitable. The term “standardized digestibility” is slowly replacing “true digestibility” in the vocabulary of animal nutritionists.
Apparent digestible energy (ADE)=IE−FE
True (or standardized) digestible energy=IE−(FE−FmE) Measurement of FmE of fish has received little attention. The FmE that has been mostly studied in fish and other animals (swine and poultry) has
been associated with endogenous protein/nitrogen losses. The most com- mon approach for measuring metabolic fecal nitrogen (MFN) representing endogenous nitrogenous losses is by determining the fecal nitrogen output of fish fed a protein-free (nitrogen-free) diet. The MFN of fish fed a protein- free diet has been estimated as about 2.7–3.3 mg/100 g live body weight per day or 123–144 mg/100 g dry diet consumed in common carp at 20◦C (Ogino et al.,1973). FmE as protein (probably contributing the most to FmE) can, therefore, be estimated to be about 0.4 kJ/100 g live body weight per day or 20 kJ/100 g dry matter intake. This is relatively small, being equivalent to about 1% of the IE or about 10–20% of the FE of animals fed good-quality practical diets.
Fish will generally eat very little of a protein-free diet, making it very difficult to calculate meaningful estimates of MFN. Moreover, there is evi- dence that the amount of MFN produced by animals receiving a semipurified protein-free diet can differ significantly from that of animals fed practical diets containing protein (Nyachotiet al.,1997). Several other dietary con- stituents (fiber, antinutritional factors) can enhance MFN (Nyachotiet al., 1997). For these reasons, it is reasonable to doubt the accuracy of “true”
protein digestibility coefficients calculated using estimates of MFN obtained from fish fed protein-free diets. Accurate estimation of MFN may require the use of sophisticated techniques (for review see Nyachotiet al., 1997).
This type of work remains to be carried out with fish.
In digestibility studies with swine and poultry, fecal samples must be coll- ected from the ileum or from cecectomized animals because of the signifi- cant activity of the intestinal microflora in the large intestine or cecum of these animals (Levis and Bayley, 1995). Reabsorption of endogenous material (e.g., enzymes) in the hindgut is thus prevented. Correction for endogenous losses is, therefore, essential to obtain the additive estimates of the apparent digestibility of nutrients for these animals. Endogenous losses from naturally voided fecal material in fish are probably small and consequently of little concern since the intestinal flora activity is generally considered negligible in most fish species (Clements, 1996) and a large proportion of endogenous material is reabsorbed prior to egestion of fe- ces. This view is supported by the higher values for the ADC of protein of most feed ingredients measured in salmonids (e.g., Cho and Bureau, 1997) compared to the ileal ADC of protein of the same ingredients in swine and poultry (Levis and Bayley, 1995).
In fish maintaining a high feed intake, the contribution of MFN to the to- tal fecal nitrogen is probably small. Under these conditions, the difference between the “true” and the apparent digestibility of protein is probably neg- ligible. If poor feed intake or poor growth is observed in a digestibility trial, it is preferable to discard the fecal samples collected since these samples may
contain a high proportion of MFN and could produce unreliable estimates of apparent digestibility (Choet al.,1982).
1.4.5. Digestibility of Whole Diets versus Digestibility of Ingredients
As discussed above, knowledge of the digestibility of energy and nutri- ents of diets is a very important aspect of any study on nutritional ener- getics. Because digestibility measurements require specialized equipment and are time-consuming, it is impossible to measure the digestibility of all diets. Because a diet is a combination of various ingredients, knowing the di- gestibility of a variety of potential fish feed ingredients may allow estimation of the digestibility of an infinite variety of diets formulated using these ingre- dients (Choet al.,1985). This, however, requires, that estimates of apparent digestibility of nutrients of different ingredients are additive, an assumption that generally holds true (Cho and Kaushik, 1990; Watanabe, 1996a,b).
Very few feed ingredients can be fed voluntarily as the sole component of a diet to fish. First, certain fish feed ingredients may not be very acceptable (palatable) for fish as a sole component of the diet. Second, it is not possible to produce feed particles with proper physical characteristics (water stabi- lity) with many individual ingredients. Third, most fish feed ingredients do not contain all the essential nutrients required by fish and feeding diets containing many of these ingredients as the sole component for more than a few days may dramatically affect the feed intake and the overall physiological status of the fish.
The use of the protocol proposed by Cho and Slinger (1979) generally solves these problems. This protocol involves comparison of the digestibility of a reference diet with that of a test diet, this test diet being a mixture of the reference diet and a test ingredient, generally at a 70 : 30 ratio. Using this protocol, palatable, water-stable, and nutritionally adequate test diets can be produced with most potential fish feed ingredients. This allows the fish to maintain a high feed intake and good growth rate, which in turn allow the measurement of apparent digestibility values that are reliable and repeatable. Also, adoption of this procedure allows the measurement of feed intake and growth rate, allowing confirmation of the nutritional adequacy of the experimental diets.
Inclusion of a digestion indicator in the reference diet allows the ADC of the energy and nutrients in the diets to be calculated from measurements of the ratios of nutrient to indicator in the diet and feces. The corresponding ADC can be calculated for the energy and nutrients in the tested ingredient by simple calculation from the ADC of the reference and test diets. The use of a reference diet, however, assumes that there are no interactions between
the components of the diet during digestion. Hence, much care is warranted in formulating such a test diet.
The apparent digestibility coefficients (ADC) for the nutrients and energy of the test and reference diets can be calculated as follows:
ADC=1−[(F/D)×(Di/Fi)] (1)
whereDis the percentage nutrient (or kJ/g gross energy) of the diet;F, the percentage nutrient (or kJ/g gross energy) of the feces; Di, the percentage digestion indicator of the diet; and Fi, the percentage digestion indicator of the feces.
The ADC of the test ingredients (ADCI) is then calculated based on the digestibility of the reference diet and the test diets as follows:
ADCI=ADCT+[((1−s)DR)/sDI] (ADCT−ADCR) (2) where ADCI is the apparent digestibility coefficient of the test ingredient;
ADCT, the apparent digestibility coefficient of the test diet; ADCR, the apparent digestibility coefficient of the reference diet; DR, the percentage nutrient (or kJ/g gross energy) of the reference diet; DI, the percentage nutrient (or kJ/g gross energy) of the test ingredient; s, the proportion of test ingredient in the test diet (i.e., 0.3); and 1−s, the proportion of reference diet in the test diet (i.e., 0.7).
1.5