Aquaculture 186 Ž2000.

311–326
www.elsevier.nlrlocateraqua-online

Replacement of fish meal in diets for Australian

silver perch, Bidyanus bidyanus
III. Digestibility and growth using meat
meal products
David A.J. Stone a , Geoff L. Allan a,) , Scott Parkinson a ,
Stuart J. Rowland b
a
NSW Fisheries, Port Stephens Research Centre, Taylors Beach NSW 2316, Australia
b
NSW Fisheries, Grafton Research Centre, PMB 3, Grafton NSW 2460, Australia

Accepted 30 November 1999

Abstract

Apparent digestibility and availability coefficients for beef and bone meal, lamb and bone
meal, a high protein meal from mixed species Žmixed meat meal, reduced ash, no bones. and from
Provine w , a high protein meal based on selected ingredients, were determined for juvenile silver
perch. Experimental diets comprised a reference diet plus meat meal products at either 15% or
30% inclusion. Silver perch readily accepted diets with up to 30% meat meal. Digestibility
coefficients for dry matter, energy, protein, and availability coefficients for amino acids were
determined to assist with the formulation of diets to assess growth of silver perch. Digestibility
coefficients for dry matter, energy, and protein all increased with increasing protein content in the
meat products. Average amino acid availability coefficients were highest for the mixed meat meal
and Provine w. Availability coefficients for alanine, arginine, glycine, methionine, proline and
serine were all significantly higher for these products than for either beef or lamb meal.
Availability of sulphur amino acids was significantly lower in Provine w than in other products.
Compared with fish meal, all meat products contained less lysine and some meat products were
also low in phenylalanine, isoleucine, and histidine. An increase in total protein content, through
removal of bone, improved the nutritional value of meat meal in silver perch diets.

)
Corresponding author. Tel.: q61-249-821232; fax: q61-249-821107.
E-mail address: allang@fisheries.nsw.gov.au ŽG.L. Allan..

0044-8486r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 4 - 8 4 8 6 Ž 9 9 . 0 0 3 8 1 - 6
312 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

Juvenile silver perch were grown for 65 days in 10 000-l tanks, using one of five diets with
similar digestible nitrogen, energy, and dry matter but different contents of fish meal, lamb meal,
and Provine w. Fish growth was reduced when diets contained less than 13% fish meal and more
than 9% Provine w. However, feed conversion efficiency and protein retention efficiency ŽPRE.
were unaffected by diet formulation. These results indicate that meat meal can replace most of the
fish meal in silver perch diets without reducing fish performance. q 2000 Elsevier Science B.V.
All rights reserved.

Keywords: Bidyanus bidyanus; Digestibility; Nutrient availability; Growth; Meat meal

1. Introduction

Fish meal replacement in aquaculture diets is recognised as a major international

research priority ŽManzi, 1989; New, 1991; Tacon, 1994. and many dietary development
studies using different species and ingredients have already been conducted ŽTacon,
1994.. The majority have investigated the potential of soybean products to replace fish
meal because of the excellent amino acid profile of soybeans Žsee review by Tacon,
1994.. Other studies have investigated a range of different products including rapeseed
meal, cottonseed meal ŽTacon, 1994., mustard oil cake, linseed and sesame meals
ŽHossain and Jauncey, 1989a,b., and other less common vegetable proteins ŽTacon,
1994..
Previous studies with meat meal products ŽLovell, 1992; Shimeno et al., 1993a,b,
1996; Tacon, 1994. have been generally positive with respect to fish meal replacement.
In Australia, meat meal and meat products may have potential to replace significant
quantities of fish meal in aquaculture feeds ŽAllan, 1997.. On a cost per unit protein
basis, meat meal is an attractive protein source for most farmed finfish and aquatic
invertebrates. Well-rendered meat meal has few anti-nutrients and does not contain
significant quantities of indigestible carbohydrates, which can restrict the use of plant
protein sources.
In Australia, meat meal is in good supply with 480 000 t of meat meal produced in
1991–1992 ŽAustralasian Agribusiness Services, 1993. and 475 000 t produced for
1995–1996 ŽPersonal communication, Graeme Banks, Australian Renderers Association,
1998..
Silver perch is an omnivorous, native Australian freshwater finfish currently being
cultured in static ponds in Australia ŽRowland and Barlow, 1991; Rowland et al., 1994,
1995.. They readily accept pelleted diets, tolerate crowded conditions, and perform well
in earthen ponds with low net water exchange ŽRowland et al., 1994, 1995..
The aim of this study was to determine the apparent digestibility coefficients ŽADCs.
for dry matter, energy, nitrogen and the apparent availability coefficients ŽAACs. for
 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326 313

2. Materials and methods

2.1. Experimental diet preparation

All ingredients were ground using a hammer mill ŽC-E Raymond, IL, USA. or sieved
to ensure all particles passed through a 710-mm screen. Dry ingredients were thoroughly
mixed in a Hobart mixer ŽTroy, OH, USA. then combined with approximately 400 ml
distilled water kgy1 dry mix before being cold pelleted through a meat mincer ŽBarnco
Australia, Leichhardt, NSW, Australia. with a 2 mm die ŽExperiment 1. or 2 and 3 mm
die ŽExperiment 2.. Pellets were dried at 358C in a convection drier for approximately 6
h until the moisture content was between 10% and 15%, to produce a dry, sinking pellet.

2.2. Experimental fish and water quality

Silver perch Ž Bidyanus bidyanus . ŽExperiment 1, mean weight" SEM: 6.2 " 0.3 g
and Experiment 2, mean weight: 12.1 " 0.1 g. were bred at the Grafton Research Centre
and raised in earthen ponds using similar techniques to those described by Thurston and
Rowland Ž1995.. Before the experiments, fish were fed SP35 ŽAllan and Rowland, 1992.
and were treated with 5 g ly1 NaCl to ensure they were free of ectoparasites and to
prevent fungal infection ŽRowland and Ingram, 1991..
Prior to stocking, fish were anaesthetised using a bath of ethyl r-aminobenzoate
ŽExperiment 1, 25 mg ly1 for 5 min or Experiment 2, 50 mg ly1 for 3 min. then caught
at random, weighed ŽExperiment 1, individually or Experiment 2, in groups of 5 or 10.
and distributed among tanks by systematic interspersion. Fish were also weighed at
harvest using the same technique.
During both experiments, water temperature Žrange 24.3–26.78C., dissolved oxygen
Žabove 5.0 mg ly1 ., pH Žbetween 7.2 and 8.3. were measured weekly using a Yeo-Kal
611 water quality analyser ŽYeo-Kal Electronics, Brookvale, Sydney, NSW Australia..
Nitrite and ammonia Ž- 0.55 mg ly1 NO 2-N ly1 and - 0.6 mg ly1 total ammonia-N
ly1 , respectively. were measured weekly using colourmetric methods described by
Major et al. Ž1972. and Dal Pont et al. Ž1973..

2.3. Biochemical analyses

All chemical analyses were done in duplicate. Feed, faecal and fish samples were
analysed for dry matter, ash, crude fat, and energy Žbomb calorimetry. by the AOAC
Ž1990. procedures. Nitrogen was determined by the method of Havilah et al. Ž1977.
Žcrude proteins N = 6.25.. Amino acids were determined by the method of Cohen et al.
Ž1989. and analysed following acid hydrolysis using high pressure liquid chromatogra-
phy and Waters Pico-Tag ŽWaters, Lane Cove, NSW, Australia.. Sulphur amino acids
were determined separately following performic acid digestion, and tryptophan, which is
lost during acid hydrolysis ŽCohen et al., 1989., was not analysed. Chromic oxide was
determined by the method described in Scott Ž1978..
314 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

2.4. Experiment 1, digestibility and aÕailability

Four meat products, beef and bone meal ŽB., lamb and bone meal ŽL., mixed meat
meal Žreduced ash, no bones. ŽM., and Provine w , were evaluated in this experiment
ŽTable 1.. The other components of experimental diets ŽTable 2. were SP35 Žat about
70% or 85% inclusion. and chromic oxide Ž1%. was used as an inert marker. SP35 was
also the control diet.
Digestibility tanks were 170-l cylindro-conical tanks Žconical base sloped at 358..
Their design and operation are described by Allan et al. Ž1999..
Fish were stocked Ž8 fish tanky1 . 7 days before faeces were collected to allow for
acclimatisation to experimental conditions. For 4 days, fish were fed SP35 and then 3
days before faeces were collected, fish were fed their respective experimental diets. Fish
were fed to excess Ž10% body weight dayy1 . using automatic conveyor belt-type feeders

Table 1
Analyses of meat meal test ingredients used in Experiment 1 Ždry basis.
Nutrient Ingredient
Beef meal Lamb meal Mixed meat Provine w Peruvian
ŽB.qbonesa ŽL.qbonesb meal ŽM. c ŽP. d fish meal e
Protein Ž%. 49.2 54.3 60.6 81.0 70.2
Fat Ž%. 9.2 7.2 17.2 10.4 11.3
Energy ŽMJ kgy1 . 16.1 16.2 23.5 25.7 20.9
Ash Ž%. 36.0 34.5 12.1 3.0 17.6

Amino acids (%)

Alanine 3.9 4.0 4.3 5.3 4.6
Arginine 3.9 4.3 4.5 6.8 5.1
Aspargine 3.3 3.8 4.9 6.9 6.2
Cystine 0.3 0.7 0.8 1.1 0.7
Glutamine 5.7 7.0 7.9 9.9 9.4
Glycine 7.7 6.7 6.6 7.9 4.8
Histidine 0.8 1.2 1.2 1.6 2.3
Isoleucine 1.3 1.8 2.1 3.7 3.5
Leucine 2.7 3.5 4.3 6.2 5.3
Lysine 2.5 3.5 3.7 4.9 5.5
Methionine 0.7 1.1 1.1 1.6 2.0
Phenylalanine 1.5 1.9 2.3 3.4 2.9
Proline 5.0 4.6 4.5 5.2 3.5
Threonine 1.6 2.1 2.6 3.6 3.2
Tyrosine 1.1 1.5 1.9 3.0 2.3
Serine 2.1 2.4 3.0 3.6 3.0
Valine 2.0 2.4 3.0 4.3 3.7
a
Supplied by Beef City, Toowoomba QLD, Australia.
b
Supplied by Fletcher International, Dubbo NSW, Australia. After lamb meal was used in the growth
experiment at GRC, Fletchers International raised the price of lamb meal to )$799rt as a new export market
was identified for this product.
c
Supplied by Midco, Maxwell NSW, Australia.
d
Supplied by Aspen Technology, Preston VIC, Australia.
e
Analysed from previous experiment.
 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326 315

Table 2
Formulation of diets used in Experiment 1 Ž% dry basis.
Ingredient Treatment
1 2 3 4 5 6 7 8 9
Reference diet ŽSP35. a 99.0 84.2 69.3 84.2 69.3 84.2 69.3 84.2 69.3
Beef meal Žwith bones. – 14.8 29.7 – – – – – –
Lamb meal Žwith bones. – – – 14.8 29.7 – – – –
Mixed meat meal – – – – – 14.8 29.7 – –
Provine w – – – – – – – 14.8 29.7
Chromic oxide 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
a
Formulation and composition of SP35 described in Table 3.

for 3 h each day from 0830 to 1130 h. Faecal samples were collected by settlement as
described by Allan et al. Ž1999..
The indirect method of Cho and Kaushik Ž1990. was used to calculate apparent
digestibility Žor availability. coefficients, with chromic oxide Ž1% dry basis. as the inert
indicator. The ADCs for energy, protein and the AACs for essential amino acids in
experimental diets were calculated as described by Cho and Kaushik Ž1990.. Then,
apparent digestibility Ž%. for dry matter, protein and energy, and apparent availability
Ž%. for amino acids for test ingredients were calculated using the methods described by
Sugiura et al. Ž1998..

2.5. Experiment 2, growth

Using the digestibility and availability data obtained from Experiment 1, five diets
with approximately the same levels of digestible energy and protein were formulated
using a linear least-cost diet formulation computer program ‘Feedmania’ ŽMania Soft-
ware, Brisbane, Australia. ŽTable 3.. The meat meal products considered for the least
cost diet formulation were Provine w , lamb meal, and beef meal. Beef meal was not
selected in the formulation by the least-cost program. The mixed meat meal product was
excluded as it is not yet commercially available. The experimental diets comprised a
reference diet ŽSP35. with 27% fish meal and three others with 13%, 6% or 0% fish
meal with the remainder replaced by a mixture of Provine w and lamb meal. For these
three diets, L-lysine, DL-methionine and L-threonine were added if necessary to adjust the
content of amino acids to that of the reference diet. The fifth diet was formulated
without fish meal or added synthetic amino acids.
Experimental tanks were 10 000-l in capacity. Recirculated freshwater was filtered
through a rapid rate sand filter before being supplied to experimental tanks at a flow rate
of 17 l miny1 . Effluent water from each tank flowed from the bottom of the tank via a
50 mm diameter drain leading to an external stand pipe. Overflow water from the
standpipes was collected and ducted into a 7000-l reservoir passed through a 1 m3
biological filter, pumped through a sand filter and then recirculated to the fish tanks.
Each tank was aerated using two air-stone diffusers.
316 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

Table 3
Formulation of SP35 ŽDiet 1. and other experimental diets to evaluate meat meal as a substitute for fish meal
in rations for silver perch ŽExperiment 2.
Ingredient Diet Ž%. Ingredient cost
1 2 3 4 5 ŽAUS$rt. a

Fish meal ŽDanish. 27.0 13.0 6.0 0.0 0.0 1300

Soybean meal 20.0 20.0 20.0 20.0 20.0 495
Blood meal 2.0 3.4 3.0 3.9 3.9 900
Lamb meal – 6.3 7.8 8.9 8.9 453
Provine w – 9.1 14.7 18.1 18.9 775
Wheat 26.9 22.1 21.9 22.0 22.2 180
Sorghum 11.0 11.0 11.0 11.0 11.0 180
Corn gluten meal 4.0 6.0 6.0 6.0 6.0 700
Millrun 2.0 2.0 2.0 2.0 2.0 190
Fish oil 1.0 1.1 1.1 1.1 1.1 800
DL-Methionine 0.2 0.3 0.4 0.5 – 5500
L-Lysine – 0.1 0.2 0.3 – 4250
L-Threonine – – 0.7 0.1 – 8500
Di-calcium phosphate 2.0 1.7 1.9 2.0 2.0 610
Vitamin premix b 1.0 1.0 1.0 1.0 1.0 4000
Mineral premix c 3.0 3.0 3.0 3.0 3.0 4000
Diet ingredient cost (AUD$r t) 756 699 727 652 607

Composition (calculated)
Digestible protein Ž%. 32.1 34.0 34.1 34.1 34.0
Digestible energy ŽMJ kgy1 . 13.0 13.4 13.3 13.2 13.3
Ash Ž%. 6.9 7.5 7.5 7.5 7.5
ADF Ž%. 3.1 3.3 3.3 3.3 3.4
Fat Ž%. 6.4 6.2 6.0 5.7 5.8
Linolenic series fatty acids Ž%. 0.3 0.3 0.3 0.3 0.3
Available arginine Ž%. 2.0 2.1 2.1 2.1 2.1
Available histidine Ž%. 0.8 0.9 0.8 0.8 0.8
Available isoleucine Ž%. 1.4 1.4 1.3 1.3 1.3
Available leucine Ž%. 2.9 3.1 3.0 3.0 3.0
Available lysine Ž%. 2.1 2.2 2.0 2.0 1.8
Available metqcys Ž%. 1.4 1.4 1.4 1.4 0.9
Available pheqtyr Ž%. 2.7 2.8 2.7 2.7 2.7
Available threonine Ž%. 1.4 1.4 1.4 1.4 1.3
Available valine Ž%. 1.7 1.7 1.6 1.6 1.6
a
Diet ingredient cost does not include the cost of processing. Prices from NSW Agriculture, Sydney Retail
Feed Ingredient Prices or from commercial feed ingredient manufacturers, 1995 Ždoes not include freight..
b
ŽIU kg y1 diet.: retinol ŽA., 8000; cholecalciferol ŽD3., 1000; a-tocopheryl acetate ŽE., 125; Žmg kg y1
diet : ascorbic acid ŽC. ŽROVIMIX STAY C 25; Roche., 1000; biotin Ž2%., 1; d-calcium panothenate, 55;
.
calcium propionate, 250; choline chloride, 1500; cyanocobalamin ŽB12., 0.02; ethoxyquin, 150; folic acid, 4;
menadione sodium bisulphite ŽK3., 16.5; myo-inositol, 600; nicotinamide, 200; pyridoxine ŽB6., 15; riboflavin
ŽB2., 25.2; thiamin HCl ŽB1., 10.
c
Žmg kg y1 diet.: calcium carbonate, 7500; manganese sulphate, 300; zinc sulphate, 700; copper sulphate,
60; ferrous sulphate, 500; sodium chloride, 7500; potassium iodate, 2.

Eighty-five silver perch were stocked ŽNovember wspringx 1995. into each of the 15
10 000-l tanks Žthree replicate tanks for each diet.. Fish were fed by hand to satiation
 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326 317

twice daily at 0800 and 1500 h, 7 days a week, for a period of 65 days. Waste feed was
minimised by observing fish during feeding and when fish lost their appetite feeding
was ceased. Fish were then harvested ŽFebruary wlate summerx 1996. and survival mean
weight increment and feed conversion ratio ŽFCR. were calculated from each tank.
Proximate analyses on the whole body composition of five fish randomly selected
from each tank were also conducted and the indices for protein deposition ŽPD., fat
deposition ŽFD., protein efficiency ratio ŽPER., protein retention efficiency ŽPRE., and
energy retention efficiency ŽERE. were calculated.

2.6. Statistical analysis

The digestibility and availability experiment ŽExperiment 1. was designed for analy-
sis using two-factor ANOVA with meat products ŽB, L, M or P. as the first factor
Žfixed. and inclusion level Ž15% or 30%. as the second factor Žalso fixed.. Single factor
ANOVA was used to assess the difference between proximate body composition of fish
fed experimental diets.
The growth experiment ŽExperiment 2. was designed for analysis using single-factor
ANOVA. Homogeneity of variances was assessed using Cochran’s Test ŽWiner, 1971.
and comparison between means were made using Student–Newman–Kuels multiple-
range test. Differences between means were considered significant at P - 0.05. Unless
otherwise stated, all results appear as mean " standard error of mean Ž n s 3..

3. Results

3.1. Experiment 1, digestibility and aÕailability

The analysed proximate composition of meat meals tested in this study is presented in
Table 1. The difference between the sum of the percentages of protein, fat and ash, and
100 for these meals indicates the presence of some nitrogen-free material. This may
indicate contamination with material Že.g., carbohydrate. from the rumen or another
extraneous source. When compared to Peruvian fish meal, the beef and lamb meals had
less protein, fat and energy; the mixed meat meal had less protein and more fat and
energy; and Provine w had more protein and energy. Provine w and the mixed meat meal
had less ash than the Peruvian fish meal, while the beef meal and lamb meal had more.
The amino acid content of the tested ingredients also differed and lysine was lower in all
meat products than in Peruvian fish meal.
Digestibility and availability coefficients for ingredients were calculated using the
values for the reference diet and the proportion of the ingredients used ŽTable 4.. There
was no interaction between ingredient or inclusion level for any nutrient Ž P ) 0.05..
There were significant differences between ingredients for all nutrients except methion-
ine Ž P - 0.05., but significant differences between inclusion levels only occurred for six
amino acids Žalanine, arginine, glycine, histidine, lysine and valine.. The average
availability coefficients Ž n s 12. for each of these six amino acids for all ingredients
included at 15% and 30%, respectively, were: alanine 77.1 " 2.6, 79.5 " 2.4; arginine
318 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

Table 4
Apparent digestibility and availability coefficients of ingredients for silver perch in Experiment 11,2,3
Ingredient Pooled SEM
Beef Lamb Mixed Provine w
Nutrient
Dry matter 42.6 a 54.9 b 75.7 c 88.9 d 3.9
Energy 71.4 a 80.8 b 81.9 b 89.9 c 1.7
Protein 69.7 a 73.4 b 83.7 c 85.5 c 1.4

Amino acids
AlaU 68.3 a 72.6 b 85.9 c 86.4 c 1.7
ArgU 72.0 a 76.0 b 87.6 c 87.6 c 1.5
Asp 77.1a 84.6 b 86.5 b 85.4 b 1.1
Cys 71.7 a 75.0 a 70.4 a 86.1a,b 1.9
Glu 75.3 a 81.6 b 86.1c 87.1c 1.1
Gly U 65.2 a 64.5a 87.8 b 90.2 b 2.6
HisU 72.0 a 87.0 b 82.9 b 84.5 b 1.4
Ile 73.2 a 83.8 b 79.6 b 80.5 b 1.0
Leu 77.0 a 84.4 b 83.8 b 82.4 b 0.8
LysU 73.6 a 83.4 b 83.8 b 84.2 b 1.0
Met 82.0 83.8 83.5 83.3 0.5
Phe 73.4 a 82.2 b 83.3 b 81.6 b 1.0
Pro 68.0 a 67.5a 86.8 b 89.0 b 2.2
Ser 72.6 a 76.8 b 85.9 c 84.4 c 1.3
Thr 74.7 a 83.3 b 84.7 b 84.3 b 1.0
Tyr 80.2 a 84.8 b 85.0 b 83.6 b 0.6
ValU 74.6 a 80.5 b 79.9 b 80.1b 1.0
Average 73.5 79.6 83.8 84.8
1
Apparent digestibility Žor availability. coefficients for test ingredients were calculated using the equation:
Apparent digestibility Žor availability. of nutrient in test ingredients Žnutrient or energy concentration in test
diet=apparent digestibility Žor availability. of nutrient or energy in test dietyproportion of reference
diet=nutrient or energy in reference diet=apparent digestibility Žor availability. of nutrient or energy in the
reference diet.rŽproportion of test ingredient in test diet=nutrient or energy concentration in test ingredient..
2
Values are means"pooled SEM; ns6 Žaverage of both inclusion levels.. There was no interaction
Ž P ) 0.05. between ingredients and inclusion level for any nutrient or amino acid. Means in the same row
with the same letter in the superscript indicate the difference between ingredients was not significant
Ž P ) 0.05..
3
Inclusion content did not affect digestibility or availability coefficients of any nutrient or those amino
acids not marked with an asterisk. For those amino acids marked with an asterisk, availability coefficients
were lower when ingredients were included at 15% compared with Ž P - 0.05..

77.6 " 2.2, 82.0 " 2.1; glycine 75.6 " 3.8, 78.2 " 3.6; histidine 79.6 " 2.3, 83.5 " 1.6;
lysine 80.0 " 1.7, 82.5 " 1.2; and valine 77.8 " 1.5, 80.6 " 1.2.
The differences in amino acid availability between the 15% and 30% inclusion levels
were small Žaverage 3.1%, range 2.4–4.4%.. In every case, the standard error was larger
for the amino acid availability at the 15% inclusion level compared to the 30% inclusion
level. The differences may be attributed to the more variable results obtained from using
the lower inclusion level and may be biologically insignificant.
For dry matter, energy and protein digestibility coefficients increased significantly
with protein content Ž P - 0.05.; Provine w was the highest, followed by the mixed meal,
Table 5
Growth performance, survival, feed utilisation, and carcass composition of silver perch after a 65-day feeding trial ŽExperiment 2. a
Diet Ž% fishmeal.
1 Ž27. 2 Ž13. 3 Ž6. 4 Ž0. 5 Ž0. b
Initial weight Žg. 11.90"0.16 12.20"0.02 12.10"0.09 12.20"0.08 12.10"0.10
Weight increment Žg. 60.40"2.29 a 60.20"1.01a 53.90"2.30 a,b 52.30"1.75 b 50.00"1.47 b
Survival Ž%. 99.20"0.40 100.00"0.00 99.60"0.40 99.60"0.40 100.00"0.00

D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

Total feed intake Žkg tanky1 . 7.61"0.26 b 7.64"0.19 b 6.43"0.41a 6.66"0.14 a 6.23"0.13 a
Feed conversion ratio c 1.48"0.01 1.44"0.02 1.46"0.02 1.50"0.02 1.47"0.02
Protein efficiency ratio d 2.10"0.02 a 2.04"0.02 a,b 2.01"0.03 a,b 1.96"0.03 b 2.01"0.03 a,b
Protein deposition Žg fishy1 dry weight. e 10.54"0.30 a 10.37"0.38 a 9.37"0.19 b 8.80"0.20 b 8.98"0.33 b
Protein retention efficiency Ž%. f 36.70"0.26 35.18"0.71 34.96"1.18 32.93"0.72 36.02"0.63
Fat deposition Žg fishy1 dry weight. g 11.02"0.76 11.37"0.38 10.21"1.20 10.14"0.86 9.42"0.23
Energy retention efficiency Ž%. h 65.32"0.79 65.80"1.82 65.97"2.05 64.94"2.38 65.99"0.76
Feed cost ŽAUS$rkg fish. i 1.12"0.01d 1.01"0.01b 1.06"0.02 c 0.98"0.01b 0.89"0.01a

Carcass composition (% dry weight basis of whole fish)

Initial
Moisture 67.37 59.67"0.28 60.00"0.33 59.61"0.48 59.99"0.44 59.58"0.04
Crude protein 42.63 41.88"0.52 41.65"0.91 41.56"1.34 40.71"1.42 42.44"0.40
Crude fat 31.37 41.91"1.24 43.54"0.88 42.71"2.46 43.95"1.66 42.44"0.40
Ash 15.40 10.61"0.39 10.76"0.03 11.08"0.52 10.90"0.39 10.79"0.13
a
Values are means"SEM for three replicate tanks. Means in rows which share the same superscript were not significantly different Ž P ) 0.05; ANOVA; SNK..
b
No added crystalline amino acids.
c
FCR s wweight of food, adjusted to 92% dry matterrwet weight fish gainx.
d
PDs wfinal weight Ždry basis.=final protein content Ždry basis.r100.yŽinitial weight Ždry basis.=initial protein content Ždry basis.r100.x.
e
FDs wfinal weight Ždry basis.=final fat content Ždry basis.r100.yŽinitial weight Ždry basis.=initial fat content Ždry basis.r100.x.
f
PER s windividual weight gain Žg.rindividual protein intake by fish Žg dry weight.x.
g
PREs wŽŽfinal dry weight=final % dry weight body protein.yŽinitial dry weight=initial % dry weight body protein..rdry weight protein intake=100x.
h
EREs wŽŽfinal dry weight=final % body energy.yŽinitial dry weight=initial % body energy..rdry weight energy intake=100x.
i
Feed costs the cost of ingredients to produce 1 kg fish ŽFCR=cost of ingredients. Žexcludes processing, handling, and transport costs; see Table 3..

319
320 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

lamb meal then beef meal. Significant differences are indicated in Table 4. The
availability of all amino acids were averaged and the values were 84.8% for Provine w ,
83.8% for the mixed meat meal, 79.6% for lamb meal, and 73.5% for beef meal.
Availability coefficients for non-essential amino acids, alanine, glycine, proline and
serine were significantly higher for the mixed meat meal and Provine w compared with
other products.

3.2. Experiment 2, growth

Diet had no significant effect on survival, FCR, PRE, FD, or ERE Ž P ) 0.05..
However, there was a significant diet effect Ž P - 0.05. on weight increment, PER and
PD, and all of these indices were lower for diets without fish meal ŽTable 5..
There were no significant differences Ž P ) 0.05. in the moisture, crude protein, crude
fat, and ash composition of silver perch carcasses fed different experimental diets for 65
days ŽTable 5.. However, there was a difference Ž P ) 0.05. between the initial and final
moisture, crude fat, and ash carcass composition for fish from each of the experimental
diets. Carcass protein composition remained unchanged.

4. Discussion

In this study, the omnivorous silver perch readily accepted diets with up to 30% meat
meal. This is consistent with other studies where diets including meat meal at contents
ranging from 30% to 70%, as a substitute for fish meal, have been readily accepted by
both omnivorous and carnivorous species such as tilapia, gilthead sea bream, rainbow
trout, and yellowtail ŽDavies et al., 1989, 1993; Watanabe and Pongmaneerat, 1991;
Shimeno et al., 1993a,b.. Mohsen and Lovell Ž1990. found that meat and bone meal at
an inclusion level of 11% increased the palatability of soybean mealrcorn based diets
for channel catfish.
Meat meals were generally well digested by silver perch with higher ADCs for dry
matter, energy and nitrogen for the meat products with less ash ŽTable 4.. Other authors
have also found protein digestibility to be negatively correlated with high ash content in
meat meals for rainbow trout ŽWatanabe and Pongmaneerat, 1991. and gilthead sea
bream ŽNengas et al., 1995..
Nengas et al. Ž1995. calculated ADCs for diets containing meat products fed to
gilthead sea bream. For a diet with 43% meat meal Ž80% protein and 4% ash., protein
and energy digestibility coefficients were 92% and 86%, respectively, while for a diet
with 40% meat and bone meal Ž52% protein and 27% ash., protein and energy
digestibility coefficients were 78% and 75%, respectively. The comparable digestibility
coefficients for protein and energy for silver perch fed diets containing Provine w Ž81%
protein and 3% ash., were 86% and 90%, respectively, and for diets with lamb meal
Ž54% protein and 35% ash. were 73% and 81%, respectively. Given the likelihood of
differences in the ingredients used in the different studies, the digestibility coefficients
are in close agreement.
 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326 321

Very little data for individual amino acid availability coefficients have been published
for warm water species. Wilson et al. Ž1981. determined coefficients for a meat and
bone meal Ž56.1% crude protein. for channel catfish and the average of amino acid
availability value was 74.3%. This is similar to averages for silver perch of 73.5% and
79.6% for beef and lamb meals, respectively, calculated in this study. Wilson et al.
Ž1981. recorded higher availability coefficients for alanine, arginine, glycine, and proline
compared with those determined for silver perch fed beef or lamb meal, while values for
silver perch were higher for aspartic acid, methionine, threonine, serine, and tyrosine.
Removal of bone Žs less ash. for the mixed species meal resulted in higher average
amino acid availability Ž83.8% compared with 79.6% and 73.5% for lamb and beef
meal, respectively.. The increased amino acid availability for the mixed species meal
may be attributed to the removal of the relatively poor quality protein from bone and
connective tissue which may not be digested as well as protein from muscle ŽNRC,
1993..
For pigs, lower and more variable lysine availability coefficients were reported for
low ash meat meal Ž50–60% protein. compared with high ash meat meal Ž43–44%
protein., and this was attributed to a higher chance of processing damage to lysine for
meals rendered without bone ŽTed Batterham, personal communication, 1993.. This
trend was not evident with silver perch for low ash meat meals in this study.
Digestibility coefficients for protein in the present study were lower than those
previously recorded with silver perch for fish meals, oilseed meals, and cereals but
similar to those recorded for grain legumes ŽAllan et al., in press.. Protein digestibility
coefficients for silver perch fed meat products compared favourably with published
values for rainbow trout. Asgard Ž1988. calculated a protein digestibility coefficient for
meat and bone meal with 51.3% protein of 59%, while Alexis et al. Ž1988. determined
protein digestibility values of 60.9% and 59.7% for defatted meat and bone meals
Ž60.2% protein, 2.5% fat, 27.2% ash and 63.1% protein, 3.5% fat, 24.4% ash, respec-
tively..
Feed conversion efficiency is influenced by dry matter digestibility. Dry matter
digestibility coefficients for the mixed meal and Provine w compared favourably with
coefficients for lower quality Peruvian fish meal, oilseed meals, and grain legumes
ŽAllan et al., in press.. Published dry matter digestibility coefficients for meat products
used in fish diets are scarce. McGoogan and Reigh Ž1996. reported a dry matter
digestibility coefficient of 64.7% for meat and bone meal which did not differ signifi-
cantly from menhaden fish meal Ž79.7%. when fed to red drum at the 30% inclusion
level. Gaylord and Gatlin Ž1996. also reported a dry matter digestibility coefficient of
86.2% for meat and bone meal, which were comparable to select menhaden fish meal
Ž93.9%.. For rainbow trout, dry matter digestibility coefficients of 43.2% and 38.8%
were determined for fat extracted meat and bone meals Ž60.2% protein, 2.5% fat, 27.2%
ash and 63.1% protein, 3.5% fat and 24.4% ash, respectively. ŽAlexis et al., 1988..
Digestibility coefficients for energy for Provine w were similar to those for Peruvian
fish meal, although energy digestibility for other meat products was lower. Digestible
energy from all meat products compared favourably with those from oilseed meals and
grain legumes ŽAllan et al., in press.. For silver perch in this study, the digestible energy
values for beef meal and lamb meal were 11.5 and 13.1 MJ kgy1 , respectively. These
322 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

are comparable with digestible energy values determined for meat and bone meal with
54.1% protein, 10.3% fat, 31.1% ash for channel catfish Ž12.26 MJ kgy1 . and rainbow
trout Ž13.33 MJ kgy1 . ŽNRC, 1993..
During this study, meat meal products were successfully used to replace half of the
fish meal in silver perch diets. Other studies have also shown that meat meal and meat
and bone meal can be successfully used to partially replace fish meal in diets for
barramundi Ž Lates calcarifer . ŽAquacop et al., 1993., gilthead sea bream ŽDavies et al.,
1989, 1993., yellowtail ŽShimeno et al., 1993a,b., and rainbow trout ŽWatanabe et al.,
1993.. Mohsen and Lovell Ž1990. found that meat meal or other animal protein meals
including fish meal improved weight gain of channel catfish fed on soybean mealrcorn
basal diets. Digestibility studies with the prawn, Penaeus monodon, also indicate
potential for meat meal with this species ŽSmith, 1995.. However, Watanabe and
Pongmaneerat Ž1991. found that growth of rainbow trout was poorer when fed with diets
where fish meal was entirely replaced with either meat meal Ž80% protein and 4% ash.
or meat and bone meal Ž52% protein and 27% ash. than when fish were fed diets
containing white or brown fish meal as the sole protein source.
In Experiment 2, growth and the amount of deposited protein declined when silver
perch were fed diets that contained less than 13% fish meal. Possible reasons for this
include: Ž1. a lower concentration of essential nutrients in diets containing less than 13%
fish meal; Ž2. reduced attractiveness or palatability of these diets, causing reduced feed
intake and hence reduced growth; or Ž3. some growth reducing compound in the diets
with less fish meal and more meat products.
As digestible crude protein and digestible energy contents were similar for all diets,
differences in proteinrenergy ratio do not account for the differences in growth.
Compared with fish meal, meat products had lower contents of lysine, methionine plus
cystine and threonine. Watanabe and Pongmaneerat Ž1991. attributed poorer perfor-
mance of rainbow trout fed meat meal based diets, compared to fish meal based diets, to
limiting amino acids in meat meal, notably lysine, methionine, and tryptophan. How-
ever, in our study, with the exception of Diet 5 Ž0% fish meal, no added amino acids.,
lysine, methionine, and threonine were balanced in all diets using crystalline amino
acids. Had amino acids been deficient, a significant difference between Diet 4 and Diet
5 would have been expected and this was not evident. It is possible that the addition of
crystalline amino acids was not effective. There is some conjecture about their efficiency
of utilisation in aquaculture diets ŽLovell, 1989; Cowey, 1992; Murai, 1992; Davies and
Morris, 1997., however, as concentrations of all three amino acids, even in Diet 5, were
above published requirements for channel catfish and tilapia ŽNRC, 1993., deficiencies
in amino acids are considered unlikely to have accounted for differences in growth. We
did not measure availability of tryptophan, although calculated contents of all diets
Ž0.37–0.48% of dry diet based on published tryptophan contents of similar ingredients to
those used; NRC, 1993., and the lowest availability coefficients of any amino acid
recorded in the present study Ž65.2%., are still above published requirements for channel
catfish, and in close agreement with requirements for Nile tilapia and common carp
ŽNRC, 1993..
In Watanabe and Pongmaneerat’s Ž1991. study, the PER, net protein utilisation, and
biological value of fish fed meat meal based diets were all lower than fish fed fish meal
 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326 323

based diets. Conversely, in our study, PER and PRE were not affected by diet. This
suggests a different response by silver perch to the meat meals used here compared with
fish species and the products used by Watanabe and Pongmaneerat Ž1991..
Total lipid content and linolenic series fatty acid levels in all diets were kept similar
with the addition of fish oil and are unlikely to account for these differences in growth.
The diet with the highest level of fish meal is likely to have contained more highly
unsaturated fatty acids ŽHUFAs. Ž0.9% of diet. than the other diets Ž0.2–0.5%..
However, it is unlikely that this difference accounted for differences in growth. In a
recent experiment, we fed silver perch a series of diets with different ratios of fatty
acids. Silver perch grew equally well on diets with negligible contents of HUFAs
Ž- 0.01% of diet. and diets with approximately 1% HUFAs Žthese diets had similar
levels of polyunsaturated fatty acids..
Feed intake was lower for fish fed Diets 3, 4 or 5 compared with those fed Diets 1 or
2. This may have been due to reduced attractiveness or palatability of the diets. In
general, meat meal and meat and bone meal have been used to increase diet attractive-
ness and or palatability ŽMohsen and Lovell, 1990; Watanabe et al., 1993.. Apart from
the control diet ŽDiet 1., the diets used in Experiment 2 contained lamb meal and
Provine w as the meat products used to replace fish meal. The amount of each of these
was determined using a least-cost diet formulation program. Nutrients were specified to
be similar to Diet 1, the fish meal content was specified at 13%, 6% or 0% and then the
mix of lamb meal and Provine w to supply these nutrients for each diet at the cheapest
price was determined. On this basis, 6.3% and 9.1% of lamb meal and Provine w were
used in Diet 2, 7.8% and 14.7% in Diet 3, 8.9% and 18.1% in Diet 4, and 8.9% and
18.9% in Diet 5. Although Provine w was highly digestible, higher contents may have
caused the diets to be less attractive or palatable or to have reduced growth. Further
investigation on Provine w as a sole ingredient at inclusion levels above 6.3% is
warranted if it is hoped that this product is to be used as a major protein source for
aquaculture feeds.
The cost of diet to produce 1 kg of silver perch was reduced by feeding diets
containing meat meal ŽTable 5.. However, when more than half of the fish meal was
replaced, a decline in growth was observed. Therefore, the added cost of the extended
grow-out period may off-set the benefits of cheaper feed costs if fish are fed diets
containing less than 13% fish meal.
Analysed body composition of silver perch fed different experimental diets indicated
they had similar protein, fat and energy contents. Similarly, Shimeno et al. Ž1993b.
reported the body composition of yellowtail fed diets containing 30% meat meal Ž65.3%
protein, 14.8% fat, and 11.4% ash. for 30 days had protein and fat contents similar to
fish fed the control diet comprised of fish meal and soybean.
Maximum inclusion levels of ingredients in formulated diets will depend not only
upon composition and digestibility but also upon the presence of anti-nutritional factors.
Meat meal can contain high contents of bone. Excess ash from bones can reduce the
physical stability of pellets, and as ash is indigestible, simply dilute the nutrient content
of meat meal based diets. However, our results show that nutrients and energy in meat
products are well digested by silver perch and meat meal can be used to replace most of
the fish meal in formulated diets without significantly reducing fish performance.
324 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

Acknowledgements

We thank Jane Frances, Rebecca Warner-Smith, and Mark Booth at Port Stephens
Research Centre and Ken O’Brien and Wendy Peasley at Wollongbar Agricultural
Institute for their excellent technical help. Helena Heasman assisted in the manuscript
preparation. This research was partially funded by the Australian Fisheries Research and
Development under the Fish Meal Replacement Sub-Program. Additional funding was
provided by the Australian Centre for International Agricultural Research, the Australian
Grains Research and Development, and the Australian Meat Research. Support from
Australian Native Fish is also gratefully acknowledged. The manuscript was critically
reviewed by Stewart Fielder and Wayne O’Connor.

References

Alexis, M., Filioglou, M., Theochari, V., 1988. Apparent digestibility measurements of feedstuffs having
potential for use in rainbow trout diets. Thalassographica 11, 19–26.
Allan, G.L., 1997. Alternative feed ingredients for intensive aquaculture. In: Recent Advances in Animal
Nutrition in Australia. University of New England, Armidale, NSW, Australia, pp. 98–109.
Allan, G.L., Parkinson, S., Frances, J., Stone, D.A.J., Booth, M.A., Rowland, S.J., Warner-Smith, R.,
Replacement of fishmeal in diets for silver perch: I. Digestibility of alternative ingredients. Aquaculture, in
press.
Allan, G.L., Rowland, S.J., 1992. Development of an experimental diet for silver perch Ž Bidyanus bidyanus ..
Austasia Aquacult. 6, 39–40.
Allan, G.L., Rowland, S.J., Parkinson, S., Stone, D.A.J., Jantrarotai, W., 1999. Nutrient digestibility for
juvenile silver perch Bidyanus bidyanus ŽMitchell.: development of methods. Aquaculture 170, 131–145.
Association of Official Analytical Chemists ŽAOAC., 1990. Official Methods of Analysis of the Association
of Official Analytical Chemists. 15th edn. AOAC, Arlington, VA, USA.
AQUACOP, Orengo, H., Cuzon, G., Thouard, E., 1993. Optimisation de l’alimentation de Lates calcarifer en
elevage. Essais de substitution de la farmine de poisson par de la farine de cretons dans la composition de
l’aliment de grossissement. In: Fish Nutrition in Practice. Kaushik, S.J., Luquet, P. ŽEds.., Coll. Les
Colloques No. 61 INRA, Paris, pp. 691–704.
Asgard, T., 1988. Nutritional value of animal protein sources for salmonids. In: Agricultural International
Conference and Exposes Congress Proceedings, Vancouver Trade and Convention Centre, Vancouver, BC,
Canada, September 6–9. pp. 411–418.
Australasian Agribusiness Services, 1993. The meat meal and tallow industry and its markets, Main Report
Prepared for the Meat Research. Australasian Agribusiness Services, Melbourne, Victoria, Australia, 94 pp.
Cho, C.Y., Kaushik, S.J., 1990. Nutritional energetics in fish: energy and protein utilisation in rainbow trout
Ž Salmo gairdneri .. World Rev. Nutr. Diet. 61, 132–172.
Cohen, S.A., Tarven, T.L., Meys, M., 1989. The pico tag method, A Manual of Advanced Techniques for
Amino Acid Analysis. Waters, Milford, MA, USA, 123 pp.
Cowey, C.B., 1992. Nutrition: estimating requirements of rainbow trout. Aquaculture 100, 177–189.
Dal Pont, J., Hogan, M., Newell, B., 1973. Laboratory techniques in marine chemistry: II. Determination of
ammonia in seawater and the preservation of samples for nitrate analysis. In: Aust. CSIRO Div. Fish.
Oceanogr. Rep. CSIRO, Sydney, 11 pp.
Davies, S.J., Morris, P.C., 1997. Influence of multiple amino acid supplementation on the performance of
rainbow trout, Oncorhynchus mykiss ŽWalbaum., fed soya based diets. Aquacult. Res. 28, 65–74.
Davies, S.J., Nengas, I., Alexis, M., 1993. Partial substitution of fish meal with different meat meal products
in diets for sea bream Ž Sparus aurata.. In: Fish Nutrition in Practice. Kaushik, S.J., Luquet, P. ŽEds.., Coll.
Les Colloques No. 61 INRA, Paris, pp. 908–911.
 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326 325

Davies, S.J., Williamson, J., Robinson, M., Bateson, R.I., 1989. Practical inclusion levels of common animal
by-products in complete diets for tilapia Ž Oreochromis mossambicus, Peters.. In: Takeda, M., Watanabe,
T. ŽEds.., The Current Status of Fish Nutrition in Aquaculture. Tokyo Univ. Fisheries, Tokyo, Japan, pp.
325–332.
Gaylord, T.G., Gatlin, D.M. III, 1996. Determination of digestibility coefficients of various feedstuffs for red
drum Ž Sciaenops ocellatus .. Aquaculture 139, 303–314.
Havilah, E.T., Wallis, D.M., Morris, R., Wollnough, J.A., 1977. A micro-colorimetric method for determina-
tion of ammonia in Kjeldahl digests with a manual spectrophotometer. Lab. Pract. 26, 545–547.
Hossain, M.A., Jauncey, K., 1989a. Studies on the protein, energy and amino acid digestibility of fish meal,
mustard oilcake, linseed and sesame meal for common carp Ž Cyprinus carpio L... Aquaculture 83, 59–72.
Hossain, M.A., Jauncey, K., 1989b. Nutritional evaluation of some Bangladeshi oilseed meals as partial
substitutes for fish meal in the diet for common carp, Cyprinus carpio L. Aquacult. Fish. Manage. 20,
255–268.
Lovell, R.T., 1989. Nutrition and Feeding of Fish. Van Nostrand Reinhold, New York, 260 pp.
Lovell, R.T., 1992. Replacing fish meal in channel catfish diets. In: Allan, G.L., Dall, W. ŽEds.., Proceedings
on Aquaculture Nutrition Workshop, Salamander Bay, 15–17 April 1991. NSW Fisheries, Brackish Water
Fish Culture Research Station, Salamander Bay, Australia, pp. 118–121.
Major, G.A., Dal Pont, J., Kyle, J., Newell, B., 1972. In: Laboratory Techniques in Marine Chemistry. A
Manual. Aust. CSIRO Div. Fish. Oceanogr. Rep. No. 51 CSIRO, Sydney, 55 pp.
Manzi, J.J., 1989. Aquaculture research priorities for the 1990s. World Aquacult. 20, 29–32.
McGoogan, B.B., Reigh, R.C., 1996. Apparent digestibility of selected ingredients in red drum Ž Sciaenops
326 D.A.J. Stone et al.r Aquaculture 186 (2000) 311–326

Tacon, A.G.J., 1994. Feed ingredients for carnivorous fish species: alternatives to fish meal and other fisheries
resources. In: FAO Fisheries Circular No. 881 FAO, Rome, 35 pp.
Thurston, S.J., Rowland, S.J., 1995. Techniques for the hatchery production of silver perch. In: Rowland, S.J.,
Bryant, C. ŽEds.., Proceedings of Silver Perch Aquaculture Workshops, Grafton and Narrandera, April
1994. Austasia Aquaculture for NSW Fisheries, pp. 29–39.
Watanabe, T., Pongmaneerat, J., 1991. Quality evaluation of some animal protein sources for rainbow trout
Oncorhynchus mykiss. Nippon Suisan Gakkaishi 57, 495–501.
Watanabe, T., Pongmaneerat, J., Sato, S., Takeuchi, T., 1993. Replacement of fish meal by alternative protein
sources in rainbow trout diets. Nippon Suisan Gakkaishi 59, 1573–1579.
Wilson, R.P., Robinson, E.H., Poe, W.E., 1981. Apparent and true availability of amino acids from common
feed ingredients for channel catfish. J. Nutr. 111, 923–929.
Winer, B.J., 1971. Statistical Principles in Experimental Design. 2nd edn. McGraw-Hill, New York, 907 pp.