Measuring Digestibility in Fish

D.P. Bureau and C.Y. Cho

Fish Nutrition Research Laboratory
Dept. of Animal and Poultry Science
University of Guelph, Guelph, Ontario, Canada, N1G 2W1

Collection of fecal material

Measurement of digestibility is an important step in the evaluation of a feed or an ingredient.

Digestibility is, in general, a good measure of the nutritive value of a feedstuff and provides a
rational basis for the formulation of diets.

There is evidence that the collection of all the feces produced by a group of fish is not a
feasible task. Digestibility measurements in fish must, therefore, rely on the collection of a
representative sample free of uneaten feed particles and the use of a digestion indicator (indirect
method). Several techniques have been used to collect fecal material from fish. Several of these
techniques, still in use for practical reasons or by preference, will generally not produce reliable
(realistic) estimates of digestibility. The most widely used of these techniques is the collection of
feces from the lower part of the intestine by stripping, suction or intestinal dissection. This
technique generally leads to an underestimation of the digestibility of nutrients (protein in particular)
because of contamination of feces with endogenous material that would otherwise be reabsorbed
before the feces are excreted (Cho et al. , 1982; Hajen et al., 1993). 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 break up of feces by fish movement may lead to
leaching of nutrients and, therefore, overestimation of digestibility of nutrients. Finally, the
technique of Smith (1971), where the feces voided into the water are collected from fish
immobilised in small metabolic chambers, yields questionable estimates of digestibility, since the
fish need to be force-fed, they often regurgitate, and show in negative protein and energy balances.

There are three methods/systems that are more likely to produce meaningful estimates of
digestibility of nutrients if used correctly: the systems of Ogino et al. (1973) in which the feces are
collected by passing the effluent water from the fish tanks through a filtration column (TUF
column), Cho et al. (1975) in which a settling column is used to separate the feces from the effluent
water (Guelph system) and Choubert et al. (1979) in which a mechanically rotating-screen is used to
filter out fecal material (St-Pee system). These three systems have been adopted in several
laboratories around the world. These two systems apparently yield very similar apparent
digestibility coefficients (ADC) of protein and lipid for several ingredients. In a study comparing
the TUF column and the Guelph system, very similar ADC of dry matter, protein, lipid and energy
were obtained with both methods for two reference diets (Satoh et al., 1992).

With the Guelph system, the fecal material is collected in a settling column as shown in
Figure 1. There are three tanks in each unit, which all drain through a common drainpipe, and a
single standpipe placed over an acrylic settling column (10 cm diameter x 40 cm high). The base of

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
1
Technical Document, University of Guelph, Ontario, Canada.
the settling column can be surrounded by a cooling jacket to minimise degradation of the fecal
material. The tanks each measure 55 cm x 40 cm and have a sloped bottom. Each tank is loaded
with 2-4 kg of fish. The velocity of the water flow is adjusted to minimise settling of the feces in the
drainpipe and maximise recovery of the feces in the settling column. Under normal operation, it has
been observed that larger feces particles are trapped in the settling column within 2 min of being
voided by the fish.

The fish are accustomed to both the tanks and the dietary regime for at least 3 days before
collection of feces begins. The fish are fed three meals daily between 0900 and 1600 hours, the diets
being offered only as long as the fish are actively feeding, to avoid wastage. One hour after the last
meal, the drainpipe and the settling column are brushed out to remove uneaten feed particles and
feces from the system. One-third of the water in the tanks is drained out to ensure that the cleaning
procedure is complete. At 0830 hours the following day, the settled feces and surrounding water are
gently withdrawn from the base of the settling column into a large centrifuge bottle by gravity.
These feces are free of uneaten feed particles and are considered to be a representative sample of the
feces produced throughout the 24-hour period. Immediately after collection of the feces, the fish are
fed again as normal, allowing repeated sampling over 6-9 days. The feces are centrifuged at 5,000 x
g for 10-20 min and the supernatant discarded. The feces are then freeze-dried and ground for
determination of the dry matter, ash, nitrogen, lipid, gross energy and digestion indicator contents.

The Guelph system allows repeated determinations, and evaluation of different diets can be
carried out at the same time as observations of apparent digestibility, growth rate, and carcass
analysis are made. A series of eight such units allows the determination of ADC for up to seven
ingredients at any one time; one unit being devoted to the reference diet. The fish are under the
rearing conditions normally applied to other nutritional experiments. The fish maintain normal feed
intake and growth rate and thus allow for the measurement of ADC values which are realistic and
applicable to a normal growing conditions. One criticism of this method is that soluble material
could be lost from the feces due to leaching. The close agreement between ADC of dry matter,
crude protein, and lipids obtained with TUF and St-Pée systems with the results obtained using the
Guelph system, however, indicate that leaching is not an important source of error. Moreover, no
significant leaching losses were observed when feces stood in the Guelph settling collector or TUF
column for 15h in freshwater (Satoh et al., 1992). The main cause of leaching is the break up of
feces particles that results from physical handling to remove from water (handling loss), which must
be avoided (Windell, 1978).

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
2
Technical Document, University of Guelph, Ontario, Canada.
Figure 1. Guelph settling column system (CYAQ-2)

Source: Cho et al. (1985).

Estimating Digestibility of Individual Feed Ingredients

Very few potentially useful feed ingredients can be fed voluntarily as the sole component of
a diet to fish. It is always necessary to combine a mixture of feed ingredients to formulate a
nutritionally balanced diet that will support good feed intake and normal physiological function.
Thus, determination of the digestibility of an ingredient requires comparison of the ADC of a
reference and a test diet, the test diet being a mixture of the reference diet and the test ingredient
(Table 1). Inclusion of a digestion indicator in the diets allows the digestibility coefficients of the
energy and nutrients in the diets to be calculated from measurement of the ratios of nutrient to
indicator in the diet and feces.

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
3
Technical Document, University of Guelph, Ontario, Canada.
Table 1. Reference diet for salmonids (Cho et al. 1985)
_________________________________________________________________

Ingredient %
_________________________________________________________________

Fish meal, herring or capelin (>68% CP, <13% ash) 30

Soybean meal (48% CP, 3% fiber) 17

Corn gluten meal (60% CP, 3% fiber) 13

Wheat middlings (17% CP, 8% fiber) 27

Vitamin premix (VIT-8404) 1

Mineral premix (MIN-8204) 1

Digestion indicator 1

Fish oil, herring or capelin 10

_________________________________________________________________

Experimental Diets
_________________________________________________________________

Reference diet Test diet

_________________________________________________________________

(%)
Reference diet 100 70

Test ingredient 0 30
_________________________________________________________________

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
4
Technical Document, University of Guelph, Ontario, Canada.
 The apparent digestibility coefficients (ADC) for the nutrients and energy of the test and
reference diets can be calculated as follows (Cho et al., 1982):

ADC= 1-(F/D x Di/Fi) (1)

where
D = % nutrient (or kJ/g gross energy) of diet
F = % nutrient (or kJ/g gross energy) of feces
Di = % digestion indicator (AIA) of diet
Fi = % digestion indicator (AIA) of feces

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 / s DI) (ADCT - ADCR) (2)

where
ADCI = Apparent digestibility coefficient of test ingredient
ADCT = Apparent digestibility coefficient of test diet
ADCR = Apparent digestibility coefficient of the reference diet
DR = % nutrient (or kJ/g gross energy) of the reference diet mash
DI = % nutrient (or kJ/g gross energy) of the test ingredient
DT = % nutrient (or kJ/g gross energy) of the test diet
s = Proportion of test ingredient in test diet mash (i.e. 0.3)

1-s = Proportion of reference diet mash in test diet mash (i.e. 0.7)

Sugiura et al. (1998)

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
5
Technical Document, University of Guelph, Ontario, Canada.
Apparent vs. true digestibility

Feces are composed of the undigested food components and the unreabsorbed 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 fecal nitrogen not directly
originating from ingested food but from the animal itself are referred to as endogenous nitrogen gut
losses (ENL). There is interest in quantifying the ENL in order to be able to calculate the “true”
digestibility of protein and amino acids in feeds and ingredients.

Salmonids will not accept (or will eat very little of) a protein-free diet making it very
difficult to calculate meaningful estimates of ENL. Moreover, there is evidence that the amount of
ENL produced by animals receiving a protein-free diet differs significantly from that of animal fed
diets containing protein. In addition, several other dietary constituents (fiber, antinutritional factors)
can enhance ENL (Nyachoti et al., 1997). For these reasons, it is reasonable to doubt of the
accuracy of “true” protein digestibility coefficients calculated using estimates of ENL obtained from
fish fed protein-free diets. Accurate estimation of ENL may require the use of more sophisticated
techniques (see Nyachoti et al., 1997 for review). This type of work remains to be carried out with
fish.

In fish maintaining a high feed intake, the contribution of ENL to total fecal nitrogen is
probably small. Under these conditions, the difference between the "true" and apparent digestibility
of protein is negligible (Hardy, 1997, Figure 2). The use of nutritionally balanced reference and test
diets (Table 1) as proposed by Cho et al. (1982) generally allows the fish to maintain high feed
intake and good growth rate, which in turn allow the measurement of apparent digestibility values
which are reliable and repeatable. Conversely, if poor feed intake or poor growth are observed in a
digestibility trial, it is preferable to discard the fecal samples collected since these samples may
contain a high proportion of ENL, and could produce unreliable estimates of apparent digestibility.

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
6
Technical Document, University of Guelph, Ontario, Canada.
Figure 2. Apparent vs "true" (N-corrected) digestibility (theoretical relationship only)

Estimates of apparent digestibility for salmonids

Table 2 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 (Table 2). 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 (Anderson et al., 1993). The
physical state of some ingredients may affect the digestibility of protein. For example, canola meal
has a rather low ADC of protein, but dehulled canola meal (canola meat) has a significantly higher
ADC for protein (Hilton and Slinger, 1986).

Table 2. Apparent digestibility coefficients of ingredients measured with rainbow trout (Cho et
al. 1982; Cho and Kaushik, 1990; Bureau et al., 199x).

Apparent digestibility coefficients (%)

Ingredients Dry Crude Lipid Energy

Matter Protein

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
7
Technical Document, University of Guelph, Ontario, Canada.
Alfalfa meal 39 87 71 43
Blood meal
ring-dried 87 85 - 86
spray-dried 91 96 - 92
flame-dried 55 16 - 50
Brewer’s dried yeast 76 91 - 77
Corn yellow 23 95 - 39
Corn gluten feed 23 92 29
Corn gluten meal 80 96 - 83
Corn distiller dried soluble 46 85 71 51
Feather meal 77 77 - 77
Fish meal, herring 85 92 97 91
Meat and bone meal 70 85 - 80
Poultry by-products meal 76 89 - 82
Rapeseed meal 35 77 - 45
Soybean, full-fat, cook. 78 96 94 85
Soybean meal, dehulled 74 96 - 75
Wheat middlings 35 92 - 46
Whey, dehydrated 97 96 - 94
Fish protein concentrate 90 95 - 94
Soy protein concentrate 77 97 - 84

References

Anderson, J.S., S.P. Lall, D.M. Anderson and M.A. McNiven. 1993. Evaluation of protein quality in
fish meals by chemical and biological assays. Aquaculture, 125: 305-325.

Bureau, D.P., A.M. Harris, and C.Y. Cho. 199x. Apparent digestibility of rendered animal protein
ingredients for rainbow trout (Oncorhynchus mykiss). Aquaculture (in press)

Cho, C.Y. and S.J. Slinger, S.J. 1979. Apparent digestibility measurement in feedstuffs for rainbow
trout. Proc. World Symp. on Finfish Nutrition and Fishfeed Technology, Hamburg, Germany. Vol.
II, p. 239-247.

Cho, C.Y., C.B. Cowey, and T. Watanabe. 1985. Finfish Nutrition in Asia. Methodological
approaches to research and development. International Development Research Centre, Ottawa.
154p.

Cho, C.Y., S.J. Slinger and H.S. Bayley. 1982. Bioenergetics of salmonid fishes: Energy intake,
expenditure and productivity. Comp. Biochem. Physiol. 73B, 25-41.

Cho C.Y. and S.J. Kaushik. 1990. Nutritional energetics in fish: energy and protein utilization
in rainbow trout (Salmo gairdneri). In: Bourne, G.H. (ed.), Aspects of Food Production,

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
8
Technical Document, University of Guelph, Ontario, Canada.
Consumption and Energy Values. World Rev. Nutr. Diet. Karger, Basel, vol.61, pp.132-172.

Choubert, G.Jr., Noue, J. de la and Luquet, P. (1979) Continuous quantitative automatic collector for
fish feces. Prog. Fish-Cult. 41, 64-67.

Hajen, W.E., R.M. Beames, D.A. Higgs, and B.S. Dosanjh. 1993b. Digestibility of various
feedstuffs by post-juvenile chinook salmon (Oncorhynchus tshawytscha) in sea water. 1.
Validation of technique. Aquaculture, 112: 321-332.

Hardy, R.W. 1997. Understanding and using digestibility coefficients in fish nutrition.
Aquaculture Magazine (May/June 1997): 84-88.

Hilton, J.W. and S.J. Slinger. 1986. Digestibility and utilization of canola meal in practical type diets
for rainbow trout (Salmo gairdnerii). Can. J. Fish. Aquat. Sci., 43: 1149-1155.

Nyachoti, C.M., C.F.M. de Lange, B.W. McBride and H. Schulze. 1997. Significance of
endogenous gut nitrogen losses in the nutrition of growing pigs: A review. Can. J. Anim. Sci., 149-
163.

Satoh, S., C.Y. Cho, and T. Watanabe. 1992. Effect of fecal retrieval timing on digestibility of
nutrients in rainbow trout diet with the Guelph and TUF Feces Collection Systems. Nippon Suisan
Gakkaishi 58, 1123-1127.

Sugiura, S.H., F.M. Dong, C.K. Rathbone, and R.W. Hardy, 1998. Apparent protein digestibility
and mineral availabilities in various feed ingredients for salmonid feeds. Aquaculture 159, 177-
202.

Smith, R.R. 1971. A method for measuring digestibility and metabolizable energy of feeds. Prog.
Fish-Cult. 33, 132-134.

Windell, J.T., J.W. Foltz, and J.A. Sarokon. 1978. Methods of fecal collection and nutrient leaching
in digestibility studies. Prog. Fish-Cult. 40, 51-55.

Bureau and Cho. 1999. Measuring Digestibility in Fish. UG/OMNR Fish Nutrition Research Laboratory
9
Technical Document, University of Guelph, Ontario, Canada.