UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effect of spoilage and processing conditions on the nutritive value of various marine protein sources… Clancy, Gordon Sean 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_fall_clancy_gordon_sean.pdf [ 8.35MB ]
JSON: 831-1.0086533.json
JSON-LD: 831-1.0086533-ld.json
RDF/XML (Pretty): 831-1.0086533-rdf.xml
RDF/JSON: 831-1.0086533-rdf.json
Turtle: 831-1.0086533-turtle.txt
N-Triples: 831-1.0086533-rdf-ntriples.txt
Original Record: 831-1.0086533-source.json
Full Text

Full Text

EFFECT OF SPOILAGE AND PROCESSING CONDITIONS ON THE NUTRITIVE VALUEOF VARIOUS MARINE PROTEIN SOURCES FOR RAINBOW TROUT (Oncorhvnchus mykiss) AND CHINOOK SALMON (Oncorhynchus tshawytscha)byGORDON SEAN CLANCYB.Sc., (Major Zoology), University of Manitoba, 1987A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of Animal ScienceWe accept this thesis as conformingto the required••THE UNIVERSITY OF BRITISH COLUMBIAJuly 1992Gordon Sean ClancyIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of Animal ScienceThe University of British ColumbiaVancouver, CanadaDate DE-6 (2/88)iiABSTRACTThis study was undertaken to assess the nutritional value offish meals (Pacific herring, Clupea harengus pallasi) and proteinhydrolyzates (ocean perch, Sebastes alutus) processed in differentways for chinook salmon (Oncorhynchus tshawytscha) in salt waterand for rainbow trout (Oncorhynchus mykiss) in fresh water. Threelevels of raw material freshness (fresh frozen, moderately spoiledand highly spoiled) and two processing temperatures (low and high)were employed in the present study.Spoilage of Pacific herring stored at 2-5°C, as determined bythe levels of total volatile nitrogen(TVN) and trimethylamine(TMA),was slow for the initial 8 days but increased rapidly until day 15.The direct distillation method with MgO yielded significantlyhigher TVN values than the microdiffusion method with K 2CO3 or KOHand the steam distillation method with NaOH. The TMA values alsovaried with the method employed. For the routine determination ofTVN and TMA microdiffusion with K 2CO3 is recommended.The levels of putrescine and cadaverine in herring alsoincreased rapidly after 8 days of storage at 2-5°C. Histaminelevels remained low (<5.0 ppm), even after 12 days of storage. Pre-freezing of the herring prior to storage at chill temperatures mayhave delayed the formation of histamine. The levels of putrescine,cadaverine, histamine, TMA and TVN were lower in the press cakemeals than in the raw material, which was possibly due to aminelosses in the press liquor and thermal degradation of amines.Elevation of processing temperature from 75°C to 100°C had littleiiieffect on spoilage indicator levels but did increase the amount ofpress liquor released from the raw materials. Progressive spoilageof the raw materials also increased the yields of press liquor andconcomitantly decreased the yields of fish meal. The levels of TVN,TMA, putrescine, histamine and especially cadaverine wereconsidered to be useful in determining spoilage of the raw materialbut only the less volatile amines (putrescine, cadaverine andhistamine) may prove useful in predicting the quality of fish meal.Spoilage of the raw material, regardless of the proteinsource, adversely affected apparent digestibility coefficients fororganic matter, gross energy and digestible energy content inchinook salmon and rainbow trout. As spoilage of the raw materialincreased, digestibility values decreased. Protein digestibilityvalues were least affected by the degree of spoilage. Since fishmeals contain high levels of protein and lipid but almost nocarbohydrate, the probable cause of reduced organic matterdigestibility of spoiled protein sources was depression of lipidavailability. Elevation of processing temperature did not adverselyaffect digestibility values in either species.Differences in digestibility values were noted between speciesfor meals produced from highly spoiled raw material and freeze-dried material. Rainbow trout seem to be more sensitive thanchinook salmon to spoilage of the raw material as digestibilityvalues of the meals produced from spoiled raw material were moreseverly depressed in rainbow trout. Lipid oxidation may haveaccounted for these differences, as chinook salmon recentlyivtransferred to sea water may be susceptible to quality of dietarylipids.Ocean perch hydrolyzates were highly digestible in bothchinook salmon and rainbow trout. Pepsin solubility was greaterthan 98% regardless of processing and spoilage conditions. The pH-stat values (utilizing enzymes extracted from rainbow trout),increased as spoilage of the raw material progressed but wereslightly reduced in the high temperature meals. The pepsinsolubility values also increased as spoilage of the herring proteinsources progressed.TVN and TMA showed little promise as methods for predictingthe nutritive value of ocean perch hydrolyzates or herring fishmeals, although monitoring TVN and TMA levels in the raw materialis recommended. However, the relationship between levels ofputrescine, cadaverine and histamine may prove useful indetermining the quality of fish meal.In conclusion, advanced spoilage of raw material (> 8 days forherring and ocean perch stored at 2-5°C) before processing resultedin depressed digestibility values in both rainbow trout and chinooksalmon.VTABLE OF CONTENTSPageABSTRACT 	  iiList of Tables 	 viiiList of Figures 	 xAcknowledgements 	  xiCHAPTER 11.0 INTRODUCTION 	 1CHAPTER 22.0 LITERATURE REVIEW 	  52.1 	 Nutritional Energetics of Salmonids 	 52.1.1 	 Energy balance of fish 	 52.1.2 	 Sources of energy 	 72.1.3 	 Comparative aspects of energy efficiency 	 92.1.4 	 Factors affecting energy requirements 	  102.1.5 	 Estimating the energy values of feed materials 122.2	Nutritive value of fish meal 	  142.2.1 	 Production of fish meal 	  172.2.2 	 Factors influencing nutritive value	Chemical composition and energy 	 Spoilage of raw material, chemicalindicators and adverse effects 	 Temperature of processing 	  422.2.3 	 Methods of measuring nutritive value 	  472.2.3.1 	 Digestibility and growth experiments(In vivo) 	  472.2.3.2 	 Enzymatic digestibility methods (in vitro) 56CHAPTER 33.0 EXPERIMENT 1. Effect of methodology on determination of totalvolatile nitrogen and trimethylamine for previously frozenPacific herring (Clupea harengus pallasi) stored at 2-5°C forup to 15 days  613.1 INTRODUCTION 	  613.2 MATERIAL AND METHODS3.2.1	 Raw material 	  643.2.2 	 Chemical analyses for TVN and TMA 	  643.2.3	 Statistical analyses 	  663.3 RESULTS AND DISCUSSION3.3.1	 Total volatile nitrogen	  663.3.2 	 Trimethylamine 	  703.4 CONCLUSIONS 	  75TABLE OF CONTENTS CONTPageCHAPTER 44.0 EXPERIMENT 2. Influence of storage and processing conditionson the levels of total volatile nitrogen, trimethylamine,putrescine, cadaverine and histamine in Pacific herring(Clupea harengus pallasi)   774.1 INTRODUCTION 	  774.2 MATERIALS AND METHODS4.2.1 	 Raw material and processing conditions 	  814.2.2 	 Chemical analyses 	  834.3 RESULTS AND DISCUSSION4.3.1 	 Influence of spoilage on indicator levels 	  874.3.2 	 Influence of thermal processing on indicators 	  944.4 CONCLUSIONS 	 103CHAPTER 55.0 EXPERIMENT 3. Influence of raw material freshness and ofprocessing conditions on the nutritive value of Pacificherring (Clupea harengus pallasi) for rainbow trout(Oncorhyncus mykiss) in fresh water and chinook salmon(Oncorhyncus tshawtscha) in salt water  1075.1 INTRODUCTION 	 1075.2 MATERIALS AND METHODS5.2.1 	 Test protein sources 	  1085.2.2 	 Diet preparation and composition 	  1105.2.3 	 Digestibility tanks and filtration 	  1175.2.4 	 Aquarium facility and experimental protocol	  1185.2.5 	 Digestibility and fish performance 	  1245.2.6 	 Pepsin and pH-stat determination	  1245.2.7 	 Statistical analysis 	  1255.3 RESULTS AND DISCUSSION5.3.1	 Chemical composition of test ingredients	  1265.3.2 	 Spoilage indicator levels 	  1325.3.3 	 Fish performance and apparent digestibilityvalues 	 1335.3.4 	 Pepsin solubility and pH-stat values	  1405.4 CONCLUSIONS 	  143viviiTABLE OF CONTENTS CONT 	PageCHAPTER 66.0 EXPERIMENT 4. Influence of raw material freshness and ofprocessing conditions on the nutritive value of ocean perchprotein hydrolyzates (Jumper Feed Systems Ltd) for rainbowtrout (Oncorhynchus mykiss) in fresh water and chinook salmon(Oncorhynchus tshawytscha) in saltwater  1466.1 INTRODUCTION 	  1466.2 MATERIALS AND METHODS6.2.1 	 Test protein sources 	  1486.2.2 	 Diet composition and formulation 	  1556.2.3 	 Aquarium facility and experimental protocol 	  1556.2.4 	 Fish husbandry and measurement of growth	  1586.2.5 	 Statistical analysis 	  1586.3 RESULTS 	  1596.4 DISCUSSION6.4.1 	 Test protein composition	  1616.4.2 	 Spoilage indicators 	  1666.4.3 	 Fish performance and digestibility	  1696.5 CONCLUSIONS 	  174CHAPTER 77.0 SUMMARY AND CONCLUSIONS 	  175REFERENCES 	  179APPENDIX 	  214viiiLIST OF TABLESTable 	 Page1. Amino acid requirements and composition of whole bodytissue of salmonids and content of herring type meals.. 	 152. Mineral content of herring meal and mineralrequirements and toxicity for salmonids 	 163. Overall averages and differences between TVN and TMAdetermination for each of seven methods 	  694. Processing and storage conditions of previously frozen(3 months at -20°C) herring 	 825. Levels of polyamines (putrescine, cadaverine, andhistamine in ppm) and volatile amines (total volatilenitrogen(TVN) and trimethylamine (TMA) in mg N/100 g)in herring prior to and after thermal processing  886. Retention of spoilage indicators(level after processing/level before processing) X 100) in the herring press cakeafter processing 	  957. Recovery of added polyamines (putrescine and cadaverine)from an extract of herring meal using gas-liquidchromatography 	  1048. Proximate compositions and energy contents of testherring meals 	  1119a. Amino acid composition of herring meal protein sources(g/100 g dry weight sample) 	  1129b. Amino acid composition of herring meal protein sources(% protein of sample) 	  11310. Mineral composition of herring test protein sources	  11411. Levels of spoilage indicators for the herring meal testprotein sources 	  11512. Composition of basal mix used for preparation of referenceand test diets (70% reference : 30% test ingredient) employedin experiments III and IV 	  11613. Experimental conditions during the rainbow trout andchinook salmon studies (herring meal) 	  12114. Yields of fish meal and press liquor as influenced byprocessing and time of storage of raw herring	  131ixLISTS OF TABLES cont....	 Page15. Initial and final weight, specific growth rate, feedefficiency and protein efficiency ratio for rainbow trout feddiets containing the herring test protein sources (30%inclusion level) for 20 days	  13416. Apparent nutrient digestibilities and digestible energyvalues (DMB) of herring test protein sources fed tochinook salmon in salt water (CS) and rainbow trout infresh water (RT) 	  13517. In vitro digestibility of herring meal protein sources.. 14118. Correlations between various digestibility values andspoilage indicators of herring test protein sources forpH-stat, pepsin solubility, organic matter, protein andgross energy digestibility with rainbow trout and chinooksalmon 	  14419. Test protein sources (Jumper Feed Systems Ltd.) 	  14920. Proximate composition and gross energy of Jumper testprotein sources 	  15121a. Amino acid composition of Jumper test protein sources(g/16 g of nitrogen) 	  15221b. Amino acid composition of Jumper test protein sources(g/100 g dry weight sample) 	  15322. Mineral compostion of Jumper test protein sources 	  15423. Experimental conditions during the rainbow trout andchinook salmon studies (Jumper feed) 	  15724. Effect of freshness of fish and processing temperatureson digestibility of Jumper test protein sources forrainbow trout and chinook salmon 	  16325. Correlations between various digestibility values andspoilage indicators of Jumper test protein sources forpH-stat, pepsin solubility, organic matter, protein andgross energy digestibility with rainbow trout and chinooksalmon 	  168xLIST OF FIGURESFigure 	 Page1. Interrelationships between nutritional factors and fishproduction in aquaculture with reference to schematicbudgeting of protein and energy flow in salmonids 	 82. Basic process for producing fish meal 	  193. Chemical indicators used to determine the freshness offish 	  284. Mean total volatile nitrogen (TVN) and trimethylamine(TMA) values, determined by seven different methods, ofpreviously frozen Pacific herring stored at 2-5°C for upto 15 days 	  675. Meal levels of polyamines (putrescine, cadaverine andhistamine) and volatile amines (total volatile nitrogen,TVN and trimethylamine by microdiffusion with K 2CO3) inthe test herring meals (as fed basis) 	  896. Gas chromatogram for standard calibration mixture ofbiogenic amines 	  1057. Pilot fish meal manufacturing machine (Chemical ResearchOrganization, Denmark) 	  1098. Mean total volatile nitrogen and trimethylamine levelsfor ocean perch test protein meals using microdiffusionwith K2CO3 	  1609. 	 Apparent organic matter and protein digestibility ofocean perch test protein sources for rainbow trout andchinook salmon 	  162xiACKNOWLEDGEMENTSThis thesis involved the work and ideas of many people. Theauthor is indebted to Richard M. Beames for his knowledge,guidance, good humour and free rein that allowed for completion ofthis thesis. The author is also greatly indebted to Dr. Dave Higgsfor his highly contagious enthusiasm for fish nutrition; also hisencouragement, guidance and open door policy that provided accessto a wealth of information. Special thanks are due to BakhshishDosanjh for his support, advice and technical knowledge.Appreciation goes out to the members of the supervisory committeefor their contribution to this thesis.The author would like to thank the following people:Elizabeth-Anne Brown for her devotion and assistance in chemicalanalysis; Roberta Massey for her keen interest and help ininterpreting the art of gas chromatography and Erin Burns-Flett forher high performance liquid chromatography analysis. The authoracknowledges Dr. Norm Haard (University of California) for pH-statanalysis and helpful suggestions.The author would also like to thank the staff of WestVancouver laboratory for their help, patience and friendship thatprovided for a very pleasurable work experience.Very special thanks are due to the author's parents and inlawsfor their support, free meals and encouragement over the years.Finally, the author will always be indebted to his lovingwife, Leslie, for without her countenance, support and friendship,the following thesis would not have materialized.1Chapter 11.0 INTRODUCTIONThe global volume of aquaculture production has increaseddramatically in the last few decades, with the production of 5.2million metric tonnes in 1975 which increased to 14.5 millionmetric tonnes in 1988 (FAO, 1990). World salmon production was277,500 metric tonnes in 1990, and this has been forecasted toincrease gradually to 350,000 metric tonnes by 1993. BritishColumbia is currently the leading producer of farmed salmon inNorth America, with production of 13,500 metric tonnes in 1990 and18,000 metric tonnes in 1991 (Chettleburgh et al., 1991; B.C.Salmon Farmers Assoc., personal comm.).The recent surge in the amount of farmed salmon appearing onworld markets, in combination with excellent wild salmon harvests,has resulted in a depressed market price for salmon. Thus, thecontrol of production costs in finfish aquaculture has become anincreasingly important issue. Since feed costs alone can accountfor up to 40-50% of the total operating expenses of salmon farms,any reduction in feed costs and/or better utilization of feed bythe fish can make the difference between financial gain orfinancial loss (Higgs, 1986; Smith, 1990).Salmon diets contain a significant amount of fish meal,conventionally ranging from 40-65% of the total composition. Todate, the use of relatively inexpensive plant protein sources hasonly resulted in replacement of a portion of the fish meal used insalmonid diets. Thus, fluctuations in fish meal prices can greatly2influence the cost of feed (Grant, 1989). Aquaculture currentlyutilizes about 10 % of the world production of fish meal (6.5million metric tonnes). This proportion is projected to increase to14% by 1992 and to 22% in the year 2000 (Kilpatrick, 1990). Totalworld production of fish meal at present seems to have reached amaximum sustainable limit. Currently, fish meal prices fluctuate inharmony with prices of soyabean and other world protein sources.Thus, increases in the price of soyabeans (ie. shortage due toenvironmental conditions) result in increases in the price of fishmeal (Theriault, 1989; Crowder, 1990). Competition with otherindustries (ie. swine, poultry, cattle and more recently pet foodmanufacturers) for fish meal has resulted in a reduction in theavailability of high quality fish meal for aquaculture. Fish mealmay be an "expensive luxury" for poultry diets but it still remainsa necessary ingredient of many aquaculture diets (Grant, 1989;Kilpatrick, 1990).To ensure an adequate supply of fish meal for an everexpanding aquaculture industry, fishmeal manufacturers are placingan increasing emphasis on quality and consistency of meal and on amore efficient utilization of the raw material. Feed manufacturersare becoming more aware of quality criteria for fish meal. Gone arethe days of purchasing fish meal for unknown growth factors, withbuyers now being more concerned with quality and nutritive value offish meals (Smith, 1990; Hardy and Masumoto, 1990).The nutritional value of fish meal varies considerably becauseof differences in the nature of the raw material (origin, season,3whole or parts of fish) and processing conditions (cooking, dryingand grinding) (Tarr, 1982; McCallum and Higgs, 1989). Thefreshness of the raw material and temperature of drying duringprocessing have been shown to greatly influence the nutritive valueof marine protein sources for salmonids, poultry and mink(Sandfeld, 1983; Pike et al., 1990; Marki, 1990). Little is known,however, of the interactive effects of the foregoing variables onthe quality of fish meal protein for salmonids.The first consideration for formulation and production ofsuccessful diets for fish is the quality and the digestibility ofthe feed ingredients and how closely the available dietary nutrientand energy levels mirror the requirements of the particular fishspecies. Diets produced from poor quality raw materials, due tospoilage of raw fish and/or suboptimal processing conditions, havean inferior nutritive value which results in adverse effects onfish health. If the raw materials are poorly digested, importantnutrients are not available for absorption and utilization by thefish. Thus, the measurement of digestibility is the first task inevaluating the potential nutritive value of a feedstuff forinclusion in a well balanced diet (Cho et al., 1982; 1985; Cho,1990).This thesis was designed to provide a better understanding offish meal quality as affected by spoilage and thermal processing ofthe raw material. The levels of various indicators of raw materialfreshness were monitored during spoilage and thermal processing.Also, the effects of spoilage and drying temperature of the raw4material on the apparent digestibility values of organic matter,protein and energy were determined in two salmonids. Various invitro enzymatic digestibility tests such as the pepsin solubilityand pH-stat test were also examined and the results were comparedwith the digestibility values obtained with chinook salmon in saltwater and rainbow trout in fresh water.This information should contribute towards providing the fishmeal industry with a means of measuring and predicting thenutritive value of fish meal without the use of expensive, longterm growth studies.5Chapter 22.0 LITERATURE REVIEW2.1 NUTRITIONAL ENERGETICS OF SALMONIDSBioenergetics is the study of the balance between energysupply in the food, energy expenditure and storage gain. Itrequires an examination of the physiological processes by whichenergy is transformed in living organisms (Cho et al., 1982; Choand Kaushik, 1990).The following information provides a brief overview of thenutritional energetics of salmonids, with emphasis on digestion.More detailed reviews on this subject have been provided byHochachka (1969), Brett and Groves (1979), Walton and Cowey (1982),Cho and Kaushik (1985; 1990), Tytler and Cowey (1985) and Cowey andWalton (1989).2.1.2 Energy balance of fishFish are complex organisms living in a continually changingenvironment. At any time the fish's energy expenditure may beeither greater or less than the energy content of its last meal.There are many places where energy is lost between intake energyand useful products but the first law of thermodynamics, theconservation of energy applies to living organisms where:IE = FE + GE + UE + ZE + HE + REIE = Is the intake of food energy and is the weight of the foodconsumed times the gross energy of a unit weight of food.FE = Is the fecal energy which is the gross energy in the feces andis the weight of the feces times the gross energy of a unitweight of feces. FE is composed of undigested food (F iE) andcompounds of metabolic origin (F mE) (ie. enzymes, sloughed6intestinal cells)GE = Gaseous energy is the energy lost in the form of gases and isrelatively small for fish but substantial for some animals(ruminants).UE = Is the total gross energy in urine which includes energy fromnon-utilized absorbed compounds from food (U iE), endproductsof endogenous origin (USE) and end products of metabolicprocesses (UmE).ZE = Is the gill excretion energy and is the gross energy in thecombustible compounds excreted through the gills.HE = Is the total heat production which is the energy lost from ananimal system in a form other than as a combustible compound.Heat production is measured by either direct or indirectcalorimetry. Heat production is measured directly by varioustemperature probes in direct calorimetry while indirectcalorimetry measures the oxygen uptake and carbon dioxideproduction using the thermal equivalent of oxygen based onrespiratory quotient (RQ) and theoretical considerations. HEis broken down into various categories such as basalmetabolism (routine and standard metabolism for fish),voluntary activity, product formation, digestion andabsorption, waste formation and excretion.RE = Is the recovered energy which is that portion of the feedenergy retained as part of the body or voided as a usefulproduct.Digestibility of a feedstuff is normally defined in terms of energyavailability where:Digestible Energy (DE) = IE - FEMetabolizable Energy (ME) = IE - (FE + UE + ZE)Recovered Energy (RE) = ME - HEBoth digestible energy and metabolizable energy values areused to determine feed quality and energy requirements.Metabolizable energy provides more information than digestibleenergy because it accounts for energy losses from the gills andurine. The bioenergetics terminology for fish may differ from7author to author but the basic concept of energy flow throughoutthe system remains the same (Fig. 1) (Jobling, 1983; Smith, 1989a).The intake energy or gross energy of the feed can be measuredby various methods. The bomb calorimeter measures the heat ofcombustion during the burning of a known amount of fuel. Thus theenergy content of the feed is measured directly. The energy contentof the feed can also be measured indirectly. During the nineteenthcentury, nutritionists working at the Weende Agricultural Stationin Germany recognized that the compounds of foods could besubdivided based on their chemistry into proximate componentsconsisting of moisture, ash, protein, lipids, carbohydrates(nitrogen free extract) and fibre. Most of the energy supply forfish is obtained from lipids, proteins and to a lesser extentcarbohydrates (Fig. 1) (Cho et al., 1982).2.1.2 Sources of EnergyVarious techniques exist for the analysis of lipid, proteinand carbohydrate from the organic matter. The average caloricconversion terms of 9.45, 5.65 and 4.10 kcal/g for lipid, proteinand carbohydrate have been determined. These values have been usedto indirectly calculate the energy content of feed and fish. Theexact caloric conversion term may vary slightly depending on thecomposition of lipid, protein and carbohydrate for the sample andon the method used to determine each proximate component. Thus fordetermination of the intake energy it is better to use a bombcalorimeter for a more accurate measure of energy (Jobling, 1983;Cho, 1992). The amount of energy available to the animal from theFEEDFormulationEnvironmentTemperatureOxygenOthersPellet abilityDurabilitySize/FinePalatabilityStabilityFEED EFFICIENCYPRODUCTSPROFITSFish QualityDIET cFINANCIALPROXIMATEANALYSES(g/kg)4602001204090100INTAKEENERGY(MJ/kg) MJ (25-35%)160-240 g0.6-1.6 MJ (3-8%)40-100 g4-8 MJ (20-40%)Feeding/HusbandryGeneticsHealthBioeng.Others8INGREDIENT QUALITY	Nutrient RequirementsPhysical Analyses 	 Nutrient BalanceChemical Analysis 	 Biological Evaluation(See below)	(Digestibility)Figure 1. Interrelationships between nutritional factors and fishproduction in aquaculture with reference to schematic budgeting ofprotein and energy flow in salmonids. B and U, branchial andurinary losses; F, fecal loss; H, heat loss; R, retention in body.The vertical bars on the left and right depict proximate and energycomponents of a typical commercial salmonid diet. Numbers inparentheses are % of total energy intake. (Modified from Cho andKaushik 1985; Cho 1990)9feed depends on the digestibility of the components.2.1.3 Comparative Aspects of Energy EfficiencyFish convert feeds into body tissue more efficiently than dofarm animals. Rainbow trout are able to convert 1.5 kg feed (highquality feed based on dry matter basis) to 1.0 kg of fish (gain ona wet weight basis) during optimal conditions (Smith, 1989a).Chinook salmon are able to convert 1.39 kg feed to 1.0 kg fish (wetweight) when water temperatures are between 9-12°C and feeding isoptimized (Kreiberg, 1991). In contrast, the most efficient warmblooded food animal (poultry) requires 2.08 kg feed to produce 1.0kg chicken (wet weight). The reason for the superior foodconversion efficiency of fish relates to the fact that they areable to assimilate diets with a higher percentage of protein,apparently because fish have a lower dietary energy requirementthan other farm animals. However, fish and poultry convert dietaryprotein to body protein at nearly the same rate (Lovell, 1989).Protein requirements of fish are higher than those of mostterrestrial vertebrates and fish preferentially utilize proteinenergy for metabolism (Smith et al., 1978). What appears to be ahigh protein requirement for fish is really a low requirement forenergy. Rainbow trout require 8.57 kcal digestible energy per gramof protein gained, while chickens, swine and beef cattle require14.75, 20.62 and 25.00 kcal digestible energy per gram of proteingained (Smith, 1989a). There are several factors that contribute tothis high energy efficiency of fish in comparison to farm animals.102.1.4 Factors affecting energy requirementsFish are poikilothermic (with the exception of tuna) withtheir metabolic rate determined by water temperature. Rainbow troutin fresh water at 7.5°C expend 11.0 kJ/kg" /day, while at 20.0°C,they expend 33.0 kJ.kg°75/day for maintenance energy. Thus as watertemperature increases, the metabolic rate of the fish alsoincreases. Farm animals are homoiothermic and they require a largeexpenditure of energy for maintenance. A chicken expends 355kJ/kg"/day at thermoneutral temperature for maintenance energy andthe rat expends 552 and 380 kJ/kg"/day at 22.0 and 30.0°C formaintenance energy (Cho and Kaushik, 1985).The effect of body weight on the metabolic rate of fish can bedescribed by the general equation Y=aW b , where Y is the metabolicrate, W is the body weight and a is the constant which is dependenton species and temperature. For rainbow trout the equation,Y=70.5W"0 was recommended by Brett and Groves (1979). For fish ingeneral the common exponent 0.855 is recommended (Brett and Groves,1979; Cui and Liu, 1990). Hence, small fish have a higher energyrequirement per unit weight than do large fish.Fish also have an advantage over farm animals in regard tosupport and travel. Fish have evolved mechanisms (i.e. swimbladder) to maintain a neutral buoyancy in the water. Thereforefish do not require the large bone structure and antigravitationalmuscles of land animals. At the same time, fish are verystreamlined and they require little energy to move quickly throughthe water (Smith, 1989a).11The excretion of waste nitrogen from protein catabolism is anenergy costly process for land animals because they concentrate andconvert ammonia to uric acid and urea for storage and excretion.Ammonia is the principal waste product when amino acids are usedfor energy. No energy is required for the conversion of protein(amino acids) into ammonia but the formation of urea from proteinrequires the expenditure of three moles of ATP per mole of ureasynthesized (Goldstein, 1982).Less energy is lost in the excreted nitrogen endproducts forfish, since ammonia has less energy per unit of nitrogen (0.79kcal/g) than does urea (5.4 kcal/g) or uric acid (8.2 kcal/g) (NRC,1981). Similarly, the net energy of retention for proteincatabolism is 4.24 for ammonia, 3.37 for urea and 2.92 kcal/g foruric acid (Smith et al., 1978). Fish also have an advantage overfarm animals since fish excrete about 85% of their waste nitrogenas ammonia down a concentration gradient (passive diffusion)through the gills into the water, with little or no energy cost.The remaining 15% is excreted as urea, uric acid, trimethylamineoxide, creatinine and creatine through the kidney and gills(Forster and Goldstein, 1969; Goldstein and Forster, 1970).Fish have to expend energy to maintain an osmotic and ionicbalance when in fresh and salt water. Fresh water fish whoseosmotic pressure of body fluid is higher than the surrounding watertend to absorb water by osmosis and lose salts. To overcome this,fresh water fish frequently excrete large amounts of water asurine. Marine fish whose osmotic pressure of body fluid is lower12than the surrounding water, tend to lose water and gain salts. Fishmake up this water loss by drinking sea water and excreting thesalts via the gut and chloride cells located in the gills. Water isabsorbed in the gut against a concentration gradient and the fishconserve water by reducing urine output. The energy required forosmotic regulation has been estimated to be about 20-30% of thefishes total energy expenditure (Eddy, 1982).Poikilothermic fish have adapted their enzymes to overcomecold temperatures in a variety of ways. These include (1)increasing the cellular enzyme concentrations, (2) changing thetypes of enzymes in the system and (3) adapting pre-existingenzymes (Haard, 1992). This has resulted in an advantage toculturing fish at cold temperatures in comparison to farm animals.Pepsins from cold water fish exhibit low Arrhenius activationenergies, high apparent Michaelis constant, low temperature optimumand high pH optima in comparison to pepsins from farm animals. Fishtrypsins and chymotrypsins also have a lower Arrhenius energy ofactivation in comparison to mammalian trypsins and chymotrypsins.Hence, enzymes from cold adapted fish species often have a higherenzymatic activity at lower temperatures than their counterpartsfrom warm blooded animals (Simpson et al., 1989; 1990;Kristijansson, 1991).2.1.5 Estimating the energy values of feed materialsThe gross energy content of a feed material can be estimatedfrom it's chemical composition or energy can be measured directlyby bomb calorimetry. The levels of lipid, protein, and carbohydrate13in the feed multiplied by the appropriate caloric factors willprovide an approximate value for gross energy. Precise estimates ofgross energy, can be obtained by bomb calorimetry which is therecommended procedure (Jobling, 1983). The gross energy level in afeed provides little information on the ability of the fish toutilize dietary energy. The next step in determining thenutritional value of a feed is to measure digestibility (Cho andSlinger, 1979; Cho et al. , 1982; 1985)Digestibility can be measured directly by the quantitativecollection of excretions or indirectly by some digestible marker inthe diet. If gill and urine excretions are collected in the directmethod, then metabolizable energy (ME) can also be calculated. Thedetermination of ME requires more time and labour to collect andanalyze the samples. Also ME values are influenced by thecomposition of the diet and the levels of feeding. Thus forpractical feed formulation with test ingredients the digestibleenergy (DE) data are probably adequate since DE values do notappear to be influenced by level of feeding and/or by environmentalfactors (Smith, 1989a; Cho and Kaushik, 1990).The relationship between DE and ME is complex in fish(Kaushik, 1989). Also, there is controversy regarding whichphysiological fuel value to use for the various species of fish andfor the various types of feeds. Therefore estimating energy valuesof feed materials from their chemical composition is notrecommended (Nose, 1989; Wilson, 1989).142.2 NUTRITIVE VALUE OF FISH MEALFish meals vary considerably in proximate composition andnutritive value. The variations are a consequence of differences inthe raw material used and in the methods of manufacturing.The species, parts of fish used (whole fish vs fish wastessuch as heads frames and offal) and season caught all influence thelevels of protein, lipid, moisture and ash in fish meal. Thestorage conditions (freshness) of the raw material andmanufacturing procedures (temperatures of cooking and drying,duration of processing, enzymatic treatment, screening andgrinding)have also been found to influence the proximatecomposition and nutritional value of the final product.Fish meal is a dry, stable powder with a high protein content(60-75%) that provides a well balanced source of essential and non-essential amino acids. It contains high levels of lysine andsulphur amino acids that are able to meet the amino acidrequirements of salmonids (Table 1). The amino acids from highquality fish meal are readily digestible and absorbed by pigs,chicken and salmonids (FAO, 1970; Pike et al., 1990). Fish meal isalso a good source of energy (high protein and lipid content butvery low carbohydrate content), vitamins and minerals (Table 2)(Windsor and Barlow, 1981; Barlow and Windsor, 1984; FAO, 1986).During the last 30 years, considerable research has been donewith respect to optimizing the yields and quality of fish meal. Thefollowing review will provide the reader with backgroundinformation on the production and utilization of fish meal as a0 •at 0rn0 enoro•-)0) • 	 -)J41301 RI 	 I:1roCO O).cc.gE -" 	 e1.0W 04 04(T1 	C0(1)COk •-)(el ••■•■••••••••(ri0404 ,41 a:l 0 44A E4-4 4i 	 Rt0 43W 4-40). $4 RIde '6) 73 f34ckou) 	 U S4 •w,row 	4-I a) 	 • 	 • 	 •7.3 ay 0 	 Cn 0) CU 	 alc c oc(I) 04 4j 	 C1)0a) 	 CU 4.) a) ea 	 • caa cn 	 >,x C 	 R. IT C.) 41 4J ..)Oa) 4-1 LC) 4-4 4-4 >1 T/cuo)-10 ,--100U0(a(el a) 01 o'P IT (1.) a) '0U u C CU73en 	 u) (1) u)	 0 (c)	, r) RI RI rt1 	 aJc a 	 a) CO CU Cacri 1:1 1:5 Ja .00 	 W 	 c (4 --CI	CO CO 	 C rt1W 	 •-4al a) (1) .0 	 .•4 >4S4 	 .41 .L.) 	 C01 	04 04 04 	 4-1 CU	>. X X X 	 a) .01.4C231:34.14-4HZO4- 'TJ N r er 	 NCOL4 C 	 •0 04.)ro 4.4Li "4a)20 0C0) 4'1 	 COIt 	 043ai 	 • E--- 0ar • al L4rn 4.) co LI-1•a) 41•-4C •p -015(7Z -$.H I-4g4 4'ea„,,,, w z=LON•VD.4,CO•In0H•01I--01•0r-N1-I r-01•LOH.0*•N01.4'•.4,0LO•N	CI 	 LO 	 •--IN CO al• • 	 •N N in1114-1•‘1,C•1CO•elVDN•NI.IC)H•H	r") 	 H	H 	 •zr• •	(,) 	 V)Z0 0 CO 01 VD 4,1 in ri 01 0 al .4. 	 ri 	 vr kip r- H o .4. 	 r.)= o al crl c.) cv m csl h .4. 4:, 	 Lr) 	 H h ■0 1-1 V 'cr 	 010 	 i....4 • • • • • • • • • • • 	 • • • • • • •C) 4u) LO In oN e-i in,--1 h N m NCO 	 C) 	 .0 .;1‘ .4. LO Hrl 	 ..4.O .z0HElI-1c110 '.0 0crkoti).4.0.4.COv:•t-. 	 .oI-1 	 ko 	 r,N 	 In 	(N3oCO■c)incoN0114Z ZH 14.LC)•N•.44•CO• • 	 •	01 	 .cr 	 co•Tr•o koZ Z 40 0 ct)U2 L-i r- H .4. o c■I ■c) to •zt. ol o) 	 co 	 co 01 1/41:) kr) rn co 	 (IN1-1-1 	i,...,,-'— —Lc')••zt.•c■•co•N•N•01•(-)•In••4, 	co	 r.)• • 	 •co•10•r-•a)•r) o• •HEl∎0 1/40 01 0 Tr,-1r- N .4, r- co 	 (Ni 	 •1' -4. -4• d' o ci 	 Lc)......o.4.o-4.,--(NI.o.4,o o H-4, 	.4, 	-1,......o.4,..--.o•1,--..o.4.I,0Xo z0....,.4..-.,h•- -„,01 	 L.0• •...„. —S-„,0 	 LO 	 1--1• • 	 •-....•Ch•-.....•N••..,„ClC1)ElZ Zl--I 4.4= KCN.,0 0 r-1 N 1-1 N 0'••■■•0S..•-I••■•••Z C.) Ci) 0 CON01001--1 N 1!) N43 • • • • • • 	 • • •Zr410 1--1 N m LC) 	 -4, 	IC) N 0 Cl(XH..-.. ..-.. ....-.... ...-.. .--... 	 ..--...01 LI-) o o o o in 	 r.) 	 o o cN) o o41 Cl ..4. .1. -v •4. Cl 	 Cl 	 cI' •zr V •t1' 	 •:1'P4 EI010H Zg El\‘zr.•-•I-0.\1/40m........01\V1/40•LO•\LIDal•.1,•\5., \\CO M 0 N• •	 • 	 •	4-1 	 e-1 	 r-1 	 4-1	d, 	C--	 0 	 r-I• • 	 • 	 •\ \\ \•4, 	 L.0	. 	 .H 0v v•cr 	 Tr• •	 1 , 	NCO • 4-1	.....-.. 	 v	‘-; 	 H• •Tr o 4-1 N V C•1 	 01 	 in ("1 I-1 N 01, .0 EnZ 0 ft cri 	 a 	 u) 	 z 	 >, 	 u) 	 a) O (a +.) 	 a) 	 o $4 1.1 a 	 )-1 	 ,-1I-i H IHI U) >1 H H •1-4 H a) >I 	 (1) 	 4 $•1 (1) 4 $4 >4 	 elZ C.) KC Kt KC C.) (.1 CD Z H 14 14 Z a. (34 U) Ei Ei HKIC 404).-IC.)>40'04)Ea)$.4ztra)IUa)zPO0:14-1IUa)tra)04.1a)4-4C.) u)r(KS$.4 oalzCowroa)tTro4-1sa0$4 00 4.144).1Z10 c0-4-43 .04.40&.4 04-4 4-4V >1W 0• Cr% 1k0 0W4-1 411V 4-1>4 • • $.4I-4 — InV RI a)COgal4-40 4-4160401=4 inczt>4 H 0 0 0E4HC..)Z0Ztor-i•en,-1•ooHXOI-44cf)A AC.,E1rdO .„. o (1) 0 0 0 0 0 LO 0 CO 0 044 OZ 0 C) • • • • • N • N • •0 0 0 n) el II) 0 0 0 VD 0 • 0 0Ei z g - 	 $-1 4-I to o al - o o o to14 H 1-4 0 4-) r) 0 r) 0 L.0 I-IH = 4 N N 1-1 co •cr0 C.) u) -°„,^44H ZC.) __E-4 0Z Z0•0a)0(00.LC)o000o0CI)E-I 1Cf/K4E)O l0 N 00l04141ZHor413”,0 E-1DI 0Z 00•0a)C.)(rlo.v.)oo0elooCOr)I•0M1tof24 H P4 0 $-i 0 I-I 0 In ,-1W E-I N4J In 0Ni-1..--..'‘11-1E-1C))Li-)CN$44140•cnZ C.) I0 CO LC) CO 0 0 LO 0 0 LO 0 0( 0 0 • ION • 0ILO0 0 •:). CO •Z H 01 en • • CO N • cN •cr • N 4-IH rx. .-1 4-4 0 CA •zr ,-) r-I ri r-I M to N(24 H N I 4-1 - N 4-i I NC414 0KC0r-I4-II-1= C14 N N •CN00 a)U)0 04- 0 0 U) 0 0g000,-I 44) fa•HWCUC04 •U)H-1 041 •.-.4 ,--) a) a) Ir/ a U) cz o o al a= tr Er, al a) - ,-1 0H 4---I S.4 XI a 0 0) C 0 4J •-I '0 CZ a;$ 4 0 0 al (a 4 0 0Cl)- ,IC.) 0 0 (..) H Z Z fa4 1:4 Cl)  m17food source for domestic animals, with emphasis on salmonidculture.2.2.1 Production of fish mealAbout one third of the annual world catch of fish is not usedfor direct human consumption but rather it is processed into fishmeal. In relation to total fish meal production, about 90% isproduced from species of fish such as anchovy, capelin and herring.The amount and type of fish meal produced is dependant upon worldfish stocks which are known to fluctuate from year to year (Windsorand Barlow, 1981).The basic price of fish meal is largely determined by theprice of competing protein sources, especially soyabeans, as fishmeal is sold on the basis of protein content. Fish meal pricesfluctuate in harmony with prices of soyabean and other worldprotein sources. Thus, an increase in the price of soyabeans (ie.shortage due to adverse environmental conditions) results in anincrease in the price of fish meal (Theriault, 1989, Crowder,1990). Total world production of fish meal (6.5 million metrictonnes) seems to have reached a maximum sustainable limit(Kilpatrick, 1990). The largest producers of fish meal are Japan,Chile, Peru, U.S.S.R., U.S.A., Norway and Denmark (FAO, 1986).The basic process of fish meal production has been reviewed byWindsor (1971), Windsor and Barlow (1981), Barlow and Windsor 1984,and FAO (1986).Fish meal is made by cooking the fish, pressing the cookedmass to remove most of the oil and a large portion of the water18(press liquor) and drying the resultant press cake to which aconcentrated fraction of the press liquor is normally added back toproduce a whole meal (Fig. 2). The fish meal is produced inspecially designed machinery for this process (FAO, 1975; Windsorand Barlow, 1981; FAO, 1986).Cooking of the raw fish occurs in a commercial cooker whichconsists essentially of a long steam-jacketed cylinder throughwhich the fish move by means of a screw conveyor. The temperatureis controlled by steam pressure and flow rate. Large fish arehashed, while smaller fish (less than 40 cm long) are fed directlyto the indirect steam cooker (FAO, 1986). Fish are heated to about75 to 100°C within 20 minutes and protein changes occur such thatcontractile protein (collagen) is coagulated, cell walls areruptured and physiologically bound water and oil are released(Windsor 1971; Barlow and Windsor 1984). The cooking temperature iscritical since without cooking the raw material can withstandconsiderable pressure without the significant release of water oroil. If the fish are incompletely cooked, the liquor cannot bepressed out satisfactorily but if the fish are over cooked thematerial becomes too soft for pressing and the increased proportionof suspended solids in the press liquor can make evaporationdifficult. Because of the variability of the raw material (natureand quality), it is difficult to generalize as to the ideal cookingconditions. However, those conditions which produce a meal with lowoil content are most desirable (Windsor and Barlow, 1981).The amount of liquor released simply by cooking can be high,se c ra en et e(c 	 n )solidsOil/stick waterEvaporatorRaw material-----,_===‘ 19  Grinder(Butch Boy) (Cooker (75 and 100°C)   Screw Press(PressliquorDryer (75 and 100°C)liPress cakesolids(Fish meal(Press cake)Fig. 2. Basic process for producing fish meal.(Experimental conditions are in brackets and represented bysolid lines while hollow lines represent further processingwhich was not done for this experiment)20often more than 60% of the weight of the raw material. Laboratorystudies have shown that increasing the cooking temperature from 60to 100°C improves the separation of water and oil during pressingfor both oily and non-oily raw materials, with the exception of codframes where there is no significant improvement of separation byincreasing the temperature from 80 to 100°C (Ward et al., 1977).The nature of the raw material greatly affects separation. Sand-eels release about 63% of their weight (total liquor), sprats about52% of their weight and also, spoilage of sand-eels (15°C for 5days) increased the release of total liquor by 5% for allprocessing temperatures employed (Ward et al., 1977). Similarstudies on spoilage of pilchards (20°C for 5 days)(Lassen et al.,1951) and sardines (10°C for 7 days)(Tozawa and Kawabata, 1987)before processing showed increased yields of press liquor anddecreased yields of fish meal.There is evidence with some oily species (e.g. capelin) thatbetter separation of liquors, especially oil, may be achieved byheating to only 50 to 60°C (Windsor and Barlow, 1981). No dryingoccurs during the cooking stage but cooking plays an important rolein sterilization and assists in ensuring freedom from pathogenicorganisms (Barlow and Windsor 1984).The pressing stage involves removal of some of the water, oiland solids. The cooked fish is conveyed through a perforated tubewhile being subjected to increasing pressure by means of a taperedshaft. The press may have a single or twin screw to preventslipping of raw material. The resulting press cake meal has a21moisture content around 55% and an oil content of 3-4% depending onthe raw material. If the cooked material is of poor quality(spoiled) then it may be difficult to press because of slipping andthe resulting press liquor may be high in solids (sludge) whichmake evaporation difficult (Windsor and Barlow, 1981).The press liquor consists of a mixture of water (78%), solids(6%) and oil (16%) which vary depending upon the nature of the rawmaterial. The press liquor is first screened to remove any solidmaterial (Fig. 2), followed by centrifugation by a decanter ordesludger. The remaining solids are then separated by centrifugalforce and the remaining oil and water are separated by disc typecentrifuges. Final clarification of the oil is achieved by furthercentrifugation with the aid of hot water (100°C). The solids may bereturned to the press cake meal to produce a whole meal (Windsorand Barlow, 1981).The water (stickwater) from the press may contain 5% of thesolids but it represents 50% of the weight of the raw material. Thestickwater can be evaporated in a series of evaporators that areheated by steam. Enzymes (proteolytic) are often used in thestickwater to prevent fouling, aid in evaporation and stabilize theviscosity of the concentrate arising from changes in the rawmaterial (Jacobsen, 1985). The resulting fish solubles may be soldto feed and/or fertilizer companies.Finally, the press cake together with the concentrated stickwater is dried from about 55% moisture content to about 10-12%. Thelow moisture content results in a stable product (in combination22with addition of antioxidant at about 200-250 mg/kg to preventlipid oxidation) that is not liable to bacterial or enzymaticdeterioration. Considerable skill is required for the optimaldrying of fish meal. If the meal is under-dried, mould or bacteriamay occur and if the meal is overdried there is a risk of scorchingthe meal. In both cases the nutritive value of the meal is reduced(Windsor and Barlow, 1981; FAO, 1986).There are two types of dryers, namely, direct heat andindirect steam. The direct dryer(flame dryer) consists of a largerotary tube in which the press cake is tumbled rapidly in a streamof very hot air (inlet 600°C). The meal reaches temperatures of 80-95°C as the rapid evaporation of water from the wet meal causescooling by the loss of heat of evaporation (see also section2.3.2.3). More than 75% of the fish meal is dried in this manner(FAO, 1986). The drying of the meal is often very rapid (10-20min.) but requires vast quantities of air and extreme careregarding air flow must be taken to prevent scorching of the meal.The indirect dryer consists of a large rotary drum with heatsupplied by steam or hot air which passes through the jacket. Thedrying temperature is controlled by the flow and pressure of steamwhich is at about 170°C. The retention time in the dryer variesdepending on the raw material but is generally 20-30 minutes(Windsor and Barlow 1981; Barlow and Windsor 1984). The meal itselfreaches temperatures of about 80-95°C and careful drying of themeal is essential for maintenance of good nutritive value. Theoptimal processing temperature for fishmeal production for23salmonids is below 90°C to ensure high protein digestibility andgood nutritive value of the meal (Pike et al., 1990).Antioxidant (i.e. ethoxyquin) is added to the meal at 200 ppmprior to drying and at 200 to 750 ppm prior to grinding tostabilize the lipid fraction of the fish meal from oxidation. Themost common method of grinding is the hammer mill which producessmall particles averaging around 40 mesh (Tyler screen). The powderis then screened and bagged for storage (FAO, 1986).Fish silage is another source of fish protein. In the silageprocess, liquefaction of whole fish or fish offal is brought aboutby enzymes already present in the fish or in the case of fermentedfish by the addition of carbohydrate (e.g. molasses) and bacterialculture. By varying the pH with formic and/or sulphuric acid andthe holding time and temperature, the mass is stabilized(preventing microbial spoilage) at about a pH of 4 for formic anda pH 2 for sulphuric acid (FAO, 1986; Jangaard, 1987). After theaddition of formic acid to the fish slurry, the mixture is heatedto about 30°C for 1 hour and then pasteurized. A decanter may beused to separate the oil from the fish silage. The fish silage maybe further concentrated by evaporation (< 50°C) using a high vacuumsystem. Antioxidant is added at this stage to prevent lipidoxidation (Jangaard, 1991). The silages are usually dark brownsemi-pastes that can be pelleted into semi-moist diets (32-38%moisture) which are fed to salmonids. The production of silage hasbeen reviewed by Taterson and Windsor (1974), Windsor and Barlow(1981), Raa and Gildberg (1982) and Jangaard (1987;1991).24An alternative to natural silaging is the production ofhydrolyzed fish products using commercial enzymes. Enzymes such aspepsin, pronase, monzym, and trypsin are used in the enzymaticproduction of fish protein concentrates. In general, coolertemperatures (40 to 60°C) are employed during production relativeto those used for the manufacture of fish meal. The pH is eitherneutral or slightly alkaline during processing.Initially, the raw material is ground up, enzymaticallydigested, pasteurized, screened (to remove excess bones anddebris), decanted (to remove some oil), thermally dried to lessthan 10% moisture and then stabilized with antioxidant. Theresulting fine, light brown powder may contain up to 35% lipid,depending upon the species and processing conditions, in contrastto 8-12% lipid in fish meal. The production of fish proteinhydrolyzates (also known as fish protein concentrate type B) hasbeen reviewed by Windsor (1969), Hale (1972), Windsor and Barlow(1981) and Sikorski and Naczk (1981).2.2.2 Factors Influencing Nutritive Value2.2.2.1 Chemical composition and energyFish meal is a brown powder which normally contains a highlevel of protein and appreciable quantities of lipid and minerals.The protein content varies with fish species; herring meals usuallycontain 68-72% protein while anchovy meals contain 65-67% protein.The protein in fish meal has a high biological value in diets foranimals. It is composed of a well balanced mixture of essential(particularly lysine and leucine) and non-essential amino acids.25The protein content of the meal does not change with overheating(scorching) but the color becomes dark brown and the product smellsburnt. The availability of the amino acids (particularly lysine) isdecreased in scorched meals (Barlow and Windsor, 1981; FAO, 1986).The level of protein decreases with increasing levels of lipid inthe fish. The level of lipid in the fish varies seasonally sincelipids are stored in the body in preparation for spawning (Conell,1975, 1990).The lipid content of fish meal ranges from 8-11% depending onthe fish species and processing conditions. Also, the fatty acidcomposition varies with fish species and water temperature. Lipidsprovide a good source of essential fatty acids as well as fatsoluble vitamins. Lipids are often stabilized with antioxidant toprevent further oxidation. All energy in fish meal comes from itsprotein and lipid content (Barlow and Windsor, 1984).The ash content of fish meal varies in relation to the rawproduct and processing conditions. If fillet scraps are used, theresulting product may be screened to remove excess scales and bonesto keep the ash content below 15%.The NRC (1981) recommends the following criteria for fishmeal: stabilized with ethoxyquin at 200 mg/kg; minimum crudeprotein 68%; crude fat range 8-11%; maximum sodium chloride 3%;maximum ash content 15%; pepsin digestibility not less than 92.5%;free from mould and salmonella (bacteria). Spoilage of raw material, chemical indicators and adverseeffectsIn comparison to other foods, fish spoils very rapidly. Thisis because of its high water content, high level of free aminoacids, high level of natural occurring bacteria and low content ofconnective tissue relative to mammalian muscle (Pedrosa-Menabritoand Regenstein, 1988). Bacterial action is the primary cause ofseafood spoilage (Shewan, 1961; 1962; Liston, 1980; Hobbs andHodgkiss, 1982; Molin et al., 1983; Hobbs, 1987; Venugopul, 1990).If fish are stored intact, the spoilage process is greatlyaccelerated, since enzymes trapped in the gut and surroundingtissue continue to function and liquefy the abdomen, causing bellyburst and solubilisation of the fish. The liquefaction of the fishincreases the cost of fish meal production and may reduce thequality of the fish meal (Aksnes, 1988; 1989).From the time of catching, fish undergo chemical changesbrought about by oxidation, by the action of bacteria present onthe surface and in the gut of the fish, as well as by the action ofenzymes within the fish (autolysis) (Pike et a/., 1990). As thefish deteriorate, protein is broken down to peptides, amines andfree amino acids which provide nutrients for the spoilage causingbacteria. Amino acid decarboxylation by bacteria is the most commonmode of synthesis of these biogenic polyamines (putrescine,cadaverine, histamine). The terms biogenic amines, polyamines andnon-volatile amines are often used interchangeably in theliterature. However, for the present study all terms refer to27putrescine, cadaverine and histamine unless stated otherwise.At the same time trimethylamine oxide (TMAO) (present in highconcentrations in marine fish) is broken down into volatile amines.These consist mainly of trimethylamine (TMA) but dimethylamine(DMA), methylamine and ammonia are also produced (Hebard et al.,1982)(refer to Experiment 1 for more information). The majorfactors leading to the formation of volatile and non-volatileamines are levels of free amino acids, bacteria, fish species andstorage time and temperature (Klausen and Lund, 1986). Thus, thelevels of volatile amines and non-volatile polyamines are useful asquality indices for decomposition of fish (Fig. 3) (Hughes, 1958;Ritchie and Mackie, 1980; Fernadez-Salguero and Mackie, 1987).The rate of volatile and non-volatile amine formation viamicrobial degradation of amino acids and proteins is highlydependant upon temperature. Storage of fish at temperatures greaterthan chill storage (0-5°) results in the rapid formation of non-volatile (i.e. putrescine, cadaverine, histamine) (Ritche andMackie, 1980; Yamanaka et al., 1986; Middlebrooks et a/., 1988;Hollingworth et al., 1990; Wendakoon et a/., 1990) and volatileamines (Hebard et a/., 1982; Fernandez-Salguero and Mackie, 1987;Veciana-Nogoues et a/., 1990). Storage at freezing temperatures (<- 6°C) however, prevents the formation of both non-volatile andvolatile amines (Mackie and Thompson, 1974; Kramer et al., 1977;Hebard et a/., 1982; Baranowski and Pan, 1985; Baranowski et a/.,1990) except for DMA which is formed in gadoid fish stored attemperatures above - 30°C but below 0°C (Castell et a/. 1971; 1973;28=•v..ICf);.*0cat /':■•1.0• CZUS 	 "0=Z 	 • 1.,... :"(CC LiiWZ 	= tz4,-=ct u) 	 c.. -cs0•ct cc(.) >- w	0 cn.= q.--.La 1— 	 ,.=v)z uizcu •=	w 	 t4-,▪ 0-6 _i CC	 cn	III Z	 cu =(/) 0 LIJ 	 -S :-	lu >-c) ,ctM LI-1 IZC aS 	 1.11 	 cu -1o0 	 ill Z 1 ...j CCI--. co_ _Li 	zZ 	=(	= 	 i• (14)	Z 1251 >• 	 z 	 1"""Z 4 .1 = CL ■C 0 ZILI ZUCC '1 < I— 	 .:48OT'• t<..J m w LL Z	 0	>-. z cc m _IoM  cc Z.; 0a.>< 	° toZ I'm 0	0 Mi...... Laj un.	LilC) M a.. i-ci)	>- 	-it =cuci) 'B',f., LLI ZU).., ... z < (/) I = ,..0-= iz w ,>-. 1. 01 0' 0 '4a0 i;■0CI 0  NZ t\fW Z •,..,= 0111- I1uul 	• ,...,.-o =	a z	cuTizt ;-•_1 0 ctr o- Go•< ‹ 	 0 	 1cc 2	=± z	 0w H' 	 ,..= -c-rsi— •	• 	 • 	 • 	 U .1=4-0< 	 reico 	 anLEI291974) (refer to Experiment 1 for more information on TVN, TMA andDMA formation ).Biogenic amine formation is related more to the activity ofmesophyllic rather than psychrophilic bacteria (Ritchie and Mackie,1980; Veciana-Nougues et al., 1990). At chill temperatures (0-5°C)Pseudomonas spp. are dominant while at elevated temperatures (10-30°C) Vibrio spp. and Photobacterium spp. are dominant (Okuzumi etal. 1990). Thus, low temperature storage of fish is very importantin controlling bacterial histamine production because, under mostconditions, little histamine formation occurs at 0 to 10°C (Kimata,1961; Hardy and Smith, 1976; Edwards and Eitenmiller, 1975; Smithet al., 1980; Behling and Taylor, 1982). However, someinvestigators have reported substantial histamine formation in fishstored at 0 to 10°C especially if the fish had been previouslysalted (Okuzumi et al., 1981; 1982; 1984; Yamanaka et al., 1984;Van Spreekens, 1987; Morii et al., 1988).The variable results obtained in these studies on theformation of histamine in fish as a function of storage temperaturemay be due to differences in the type and level of bacterial florapresent in the fish for the various studies. Histamine producingbacterial species and strains vary considerably in their optimalconditions for growth. Histamine may not be formed in fish withhigh levels of free histidine when stored at elevated temperaturesif the predominant bacteria are not histamine producers (Edmundsand Eitenmiller, 1975; Behling and Taylor, 1982; Taylor, 1986).Freezing fish before chill storage (0-5°C) can result in an30increased lag phase in the production of TVN and TMA (Luijpen 1958;Shewan 1961). A similar delay in the production of histamine inpreviously frozen fish stored at chill temperatures has also beenreported (Baranowski and Pan, 1985; Baranowski et al., 1990). Thephotobacteria responsible for the production of histamine at chilltemperatures are sensitive to freezing. Bacteria numbers werereported to decrease from 10 6-107 per g down to less than onepercent (104-10 5 bacteria per g) in herring fillets after storage at-25°C for 2.5 months (Van Spreekens, 1987). An increased lag phasefor the production of putrescine and cadaverine in previouslyfrozen fish stored at chill temperatures is possible if bacterialpopulations responsible for the production of putrescine andcadaverine are reduced during frozen storage. Few investigatorshave studied the formation of volatile and non-volatile amines inthawed fish (Gill and Thompson, 1984; Farn and Sims, 1987).Polyamines (putrescine, cadaverine, spermidine and spermine)are naturally occurring compounds (low cellular concentrations aretightly regulated) in animal tissue. These compounds are importantin cell growth and differentiation and have been implicated inprocesses such as stimulation of DNA, RNA and protein synthesis.Their polycationic structure (positive charges distributed at fixedlengths along a conformationally flexible carbon chain) allowspecific interactions with cell membranes and therefore polyaminesinfluence and modulate membrane properties and functions (Pegg andMcCann, 1982; Bachrach et a/., 1983; Tabor and Tabor, 1984; Seiler,1990).31The membrane of the small intestine is capable of remarkableadaptive changes in its mucosal structure and function. Changes inpolyamine synthesis may be the key intracellular event initiatingthese adaptive changes. Polyamines are necessary for mucosalgrowth, development, maturation, adaptation and recovery frominjury of the small intestine. Thus, polyamines have an importantrole in the protection of cells from deleterious environments bymodification of cell membranes (Hosomi et al., 1987; Mizui et al.,1987; Schuber, 1989; Seiler, 1990).The beneficial effect of orally delivered putrescine inovercoming the antinutritional (reduced digestion and absorption ofnutrients) factors present in soyabean meal has been reported forcalves (Grant et a/., 1989) and pigs (Grant et a/., 1990). Thefeeding of diets with putrescine at low concentrations (0.2% dietor 2000 ppm) has been found to increase the growth of chickens by30%. However, putrescine was found to reduce growth and inducetoxicity at 0.8% and 1.0% of the diet (Smith, 1990b). In contrast,Colnago and Jensen (1992) reported no increase in growth whenchickens were fed putrescine at 0.05 and 0.1% of the diet. Similarresults were found for rainbow trout, as putrescine added at 0.1 to0.4% to diets containing soyabean and corn gluten meals had noeffect on feed intake and growth rate over a 12 week- period(Cowey, 1991). Further studies with rainbow trout revealed that adietary putrescine level of 1.3% is toxic and resulted in reducedgrowth rate and feed intake (Cowey, 1992- pers. comm.).Dietary TMA has been shown to extend the survival time of32guppies transferred directly from fresh water to salt water by 50%.Moreover, TMA has been shown to activate enzyme systems implicatedin water and salt permeability in the gills and digestive tract ofthe guppy, which may explain the increased survival (Daikoku 1977and Daikoku, 1980; cited in Daikoku et al., 1987). Dietary TMA at170 ppm was found to increase the growth rate and feed intake ofchum salmon in both fresh and salt water but dietary TMA at 1710ppm decreased the growth rate and feed intake of chum salmon inboth fresh and salt water (Daikoku et al., 1987). Dietary TMA at1.0% of the diet has been found to be a feeding stimulant forrainbow trout but a feeding deterrent for Atlantic salmon (Mearnset al., 1987) and chinook salmon (Hughes, 1991). The effect in thelatter two salmon species may explain the reduced feed intake andgrowth rate of the chum salmon, reported above. Also, it isnoteworthy that a combination of betaine (an appetite stimulant)and TMA in the feed of chinook salmon did not result in theaversion of the diet (Hughes, 1991).Of the biogenic non-volatile amines studied, histamine hasreceived the most attention due to its toxicological effects(scombridae poisoning or histaminic intoxication) (Arnold andBrown, 1978). Histamine is a primary amine arising from thedecarboxylation of free histidine in fish tissues (muscle). It mayexist in a variety of different forms at physiological pH becauseof its various states of ionization, tautomeric properties and thevarious conformations of the side chain. Endogenous histamine playsimportant roles in several normal and abnormal biological processes33including vasodilation, anaphylaxis and gastric acid secretion(Douglas, 1985; Taylor, 1986).Scombroid poisoning usually involves consumption of fish whichcontain large amounts of free histidine in their muscle tissue.Thus fish from the families Scombridae (tuna, mackerel, bonito,cero and sierra) and Scomberesocidae (saury) have been found tohave high levels of free histidine and are most commonly implicatedin histamine poisoning. However, non-scombroid fish belonging to■the families Pomatomidae (bluefish), Coryphenaenidae (mahi-mahi),Corangidae (jack mackerel, amberjacks, yellow tail), Clupeidae(herring, sardine) and Engranlidae (anchovies) have been implicatedin histamine poisoning as well (Taylor, 1986). The levels of freehistidine in the muscle of fish are known to vary between speciesas well as seasonally (Arnold and Brown, 1978; Pan, 1985). Freehistidine content in herring were found to vary with season from260 to 1,600 ppm with the highest levels reported in the summer(Hughes, 1959).The symptoms of scombroid poisoning (i.e. gastrointestinal-nausea, vomiting, diarrhea, abdominal cramps; cutaneous-rash,urticaria, edema; hemodynamic-hypotension and neurological-flushing, itching, burning, tingling, headache) are the same asthose for histamine poisoning (Murray et al., 1982).Antihistamines, H I antagonists (diphenhydramine or chlorpheniramine)or H2 antagonists (cimetidine) are usually selected for treatment.However, histamine does not appear to be the sole causative agentsince histamine by itself is not toxic when taken orally (Douglas,341985; Taylor, 1986).The synergistic or potentiating effects of other amines(trimethylamine, putrescine, and cadaverine) in combination withhistamine have been reported. Hayashi (1954, as cited in Arnold andBrown 1978) found that TMAO, TMA, agmatine and choline workedsynergistically with histamine in causing contractions of guineapig uterus. Alternatively, Kawabata et al. (1956) disputed thisobservation and found that TMA and TMAO were ineffective inpotentiating the action of histamine on guinea pig uterus. Parrotand Nicot (1965 as cited by Arnold and Brown, 1978) were the firstto demonstrate that the oral toxicity of histamine could bepotentiated when large concentrations of putrescine (2000 ppm) werefed to guinea pigs. Further, Bjeldanes et al. (1978) reported thatoral toxicity of histamine was greatly increased when cadaverinewas fed in combination with histamine to guinea pigs. Both lowlevels of cadaverine (<150 ppm) in the presence of histamine (150ppm) and high levels of cadaverine (150 ppm) in the presence of lowlevels of histamine (50 ppm) fed orally to guinea pigs resulted inincreased toxicity, especially when cadaverine was fed 40 min.before histamine.Cadaverine has been shown to increase the rate of transport ofhistamine across the intestinal wall of the guinea pig. This wouldresult in a higher concentration of absorbed intact histaminewithin the body (Jung and Bjeldanes, 1979). Both putrescine andcadaverine are able to bind with pig gastric mucin in vitro andtherefore they are able to liberate histamine from intestinal35mucin. This would increase the available amount of histamine thatcould be absorbed in the intestinal lumen and could result inhistamine toxicity (Chu and Bjeldanes, 1981). Putrescine andcadaverine also inhibit the histamine-metabolizing enzymesdiamineoxidase (DAO) and histamine N-methyltransferase (HMT) whichleads to increased uptake of unmetabolized histamine from the ratintestinal lumen in vitro (Taylor and Leiber, 1979). Thusputrescine and cadaverine may potentiate the toxicity of histamineby inhibiting its detoxification by DAO and HMT in vivo (Taylor andSumner, 1987).Blonz and Olcott (1978a) demonstrated that spoiled tunacontaining histamine (2500 ppm) caused regurgitation in pigs,especially those weighing 18 to 36 kg as compared to those weighing50 to 73 kg, but not in cats, dogs and rabbits. Blonz and Olcott(1978b) reported that weight gains of quail were depressed whenthey were fed dried spoiled tuna. Otake et a/. (1977,1979) foundthat the administration of histamine to mackerel via the diet orinjection into the stomach caused activation of pepsinogen topepsin, increased stomach acidity and increased cathepsin Dactivities in the spleen which resulted in stomach erosion andulceration. Watanabe et al. (1987) confirmed these findings inrainbow trout and chicks which were fed diets containing histamineor heated fish meals in combination with histamine and histidine.Fish fed diets containing histamine at either 500 ppm or 1000 ppmdeveloped a 33% and 80% reduction of gastric folds in the stomachafter 10 weeks of feeding. Chicks fed the same diets showed low36feed consumption and high gizzard erosion (GE) scores and one chickwas dead after 7 days when fed the diet containing 1000 ppmhistamine. High dietary levels of histamine (2000 ppm) are requiredto induce appetite loss and lethargic movement of tilapia (Pan,1988). Similarly, diets with high histamine contents (438 ppm) haveresulted in increased mortality (70-80%) of rainbow trout fry(Teskeredzic et al., 1989).High levels of histamine have been found in fish mealsproduced from spoiled raw materials, especially when the fishsolubles have been incorporated during meal production. When thesemeals were fed to chicks a high incidence of GE was observed(Toyama et al., 1981; 1985). The optimum temperature for histidinedecarboxylase, the enzyme responsible for histamine formation infish muscle, has been found to be 55°C, while the optimumtemperature for fish muscle protease was noted to be 60°C (Pan,1985). Eitenmiller et a/. (1982) reported a lower optimumtemperature for histidine decarboxylase, namely 37°C. Also, theyfound that the production of histidine decarboxylase was dependantupon incubation temperature, culture age, pH and free histidineconcentration with optima varying between species and strains ofbacteria.Generally, an increase in processing temperature and/or anincrease in the duration of thermal processing during fishprocessing results in a reduction of the initial histamine level.Similarly, the higher the pH and/or lower the water activity, thegreater the loss of histamine associated with processing due to the37non-enzymatic browning reaction (Maillard reaction) as well ashistamine binding with proteins (Toyama et al., 1982; Tanaka etal., 1986). A similar scenario occurs for putrescine andcadaverine, as thermal processing decreases their levels in fish(Farn and Simms, 1987). By contrast, thermal processing increasesthe levels of TMA and DMA in fish because TMAO is heat sensitive(Sigurdsson, 1947; Hughes, 1958, 1959; Tokunaga, 1975; Hebard etal., 1982; Gallardo et al., 1990).Although the initial levels of histamine decrease duringthermal processing of fish meal, the incidence of GE in chicks hasbeen shown to increase with the use of the heated fish meals. Thusthe resultant substances produced during fish meal processing maycause GE (Toyama et al., 1982; Tanaka et al., 1986; Wessels andPost, 1989).The causative substance found in some fish meals for GE and"black vomit" in chicks has recently been identified as 2-amino-9(4-imidazolyl)-7-azanonanic acid (gizzerosine). The compound is atoxic derivative of histamine or histidine (Okazaki et al., 1983).Gizzerosine is found in overheated (>130°C) fish meals and has beenshown to stimulate gastric acid secretion in chicks. The effect ofgizzerosine is depressed by cimetidine, an antagonist of thehistamine H2 receptor. Gizzerosine is about 10 times as active ashistamine in stimulating gastric secretion in vivo (Masumura eta/., 1985). The potency of gizzerosine has been found to beapproximately 1,000 fold higher than that of histamine inincreasing intracellular cyclic adenosine-3' , 5 ' -monophospate levels38in chicken proventriculus. Also, gizzerosine remains in the bloodof chickens far longer than histamine, which might explain the highincidence of GE in chicks fed gizzerosine compared to those fedhistamine (Ito et al., 1988).The high mineral content of fish meals (> 15%) has been shownto be a potential inducing factor for gastric histamine metabolismand secretion in the rat (Cheu and Fisher 1983). Rats fed dietswith menhaden fish meal have exhibited increases in stomachhistidine decarboxylase (HDC) activity as well as acid secretionwhen compared to casein fed controls. The addition of calcium tothe casein diet (same concentration as found in the fish meal),also elevated the levels of gastric histamine. Thus, Ca can triggerthe release of gastrin, which results in the release of stomachhistamine and/or facilitation of mast cell histidine incorporationwith subsequent histamine synthesis (Cheu and Fisher, 1983).The U.S. Food and Drug Administration has recently (1982)established a hazard action level for histamine in tuna of 500 ppm(Taylor, 1986) while Sweden (1980) has established a maximumtolerance of 200 ppm of histamine in fish products (Murray et al.,1982). Both Arnold and Brown (1978) and Taylor (1986) recommendfurther examination of the identity, levels and potency ofpotentiators (putrescine, cadaverine, and possibly others?) beforestrict regulatory limits for histamine in foods are imposed.Spoilage of lipid also occurs in fish during storage and therate of spoilage is also dependent upon storage temperature, oxygenlevels, and fish species. Fatty fish tend to spoil faster than lean39fish and the greater the temperature the faster the rate of lipidoxidation. Lipid oxidation also occurs in frozen fish but theprocess is reduced as the holding temperature decreases.Spoilage may occur by (1) physicochemical means due to theabsorption of odoriferous substances, (2) biochemically due to theaction of tissue enzymes and micro-organisms and (3) auto-oxidatively as a result of the oxidation of fatty acids byatmospheric oxygen. With respect to the foregoing processesinvolved, autoxidation is the most important.Unsaturated fatty acids react spontaneously with atmosphericoxygen and as the number of double bonds increase the tendency foroxidation is enhanced. Poly-unsaturated fatty acids (PUFA) (highcontent in fish) are very susceptible to auto-oxidation. Theprocess begins with the splitting of H atoms (free radicals) andthe formation of hydroperoxides and the reaction proceeds rapidlyas a chain reaction. The initiation of oxidation of PUFA isaccelerated by heat, light, metals and enzymes. The hydroperoxidesformed can react with oxygen to form secondary products such asepoxy-, keto-, dihydroperoxides, cyclic peroxides and bicyclicendoperoxides. These secondary products decompose to form alkanol,alkenal, alkenone, alkadienal and alkantrienal or they may condenseinto dimers and polymers. These products can interact withbiological molecules in fish such as pigments, enzymes, proteins,membranes and DNA. This results in discoloration, flavour and odourdeterioration, texture loss, loss of water holding capacity andvitamins. Comprehensive reviews of the mechanisms of lipid spoilage40have been presented by Hardy (1980), Khayat and Schwall (1983) andHsieh and Kinsella (1989).The measurement of oxidation in fish products presents a majorproblem that has not been resolved satisfactorily. Variouscompounds are formed during lipid oxidation such as ketones,aldehydes, alcohols, hydrocarbons, acids and epoxides. However,since endproducts are continually being lost to further reaction,measurement of the degree of oxidation is very difficult. The fivemost popular laboratory methods for determining lipid oxidation arethe free fatty acid (FFA) content, 2-thiobarbituric acid test(TBA), peroxide value (PV), carbonyl value (CV) and anisidinevalue. All methods suffer from the transitory nature of theproducts as initially they increase to a maximum and then decline(Barlow and Windsor, 1981; Woyewoda et al., 1986).Dosanjh et a/. (1988) reported that there may be considerablevariation in the quality of marine lipids due to freshness of fishbefore processing, processing conditions, levels and types ofantioxidants used, storage and handling conditions of the oil afterprocessing which all influence the degree of oxidative rancidity.Oxidized oils in animal diets may exert adverse nutritionaleffects. Severe oxidation of the lipids may for example, causetoxicity and even death if ingested (Wiseman, 1986). Moreover,lipid oxidation may reduce the digestibility of dietary lipids inchickens (March et a/., 1965; Obstvedt, 1973a, b), rainbow trout(Watanabe et a/., 1983), Atlantic salmon (Austreng and Gjefsen,1981), and carp (Iijima et a/., 1983). Protein digestibility,41however, does not appear to be affected adversely by dietaryoxidized lipid in chickens, rats, mink (Njaa et al., 1980), rainbowtrout (Watanabe et al., 1983) and Atlantic salmon (Austreng andGjefsen, 1981).However, oxidized lipid has been found to inactivate ordiminish the activity of trypsin, pepsin, alpha-chymotrypsin, andsuccinate dehydrogenase in vitro (Roubal and Tappell 1966; Hata andKanida 1980). Therefore it is not known if protein digestibilitywould be adversely affected by lipid oxidation in chinook salmon.Iijima et al. (1983) reported that oxidized lipids did notaffect the conversion of glycerol-3 phosphate into phosphatidicacid but they did inhibit the conversion of phosphatidic acid todiglyceride and diglyceride to triglyceride in carp intestine. Thetoxic effects of dietary lipid peroxides are mainly caused bydecomposition products because peroxides are converted into hydroxyor oxo-derivatives in the carp intestine prior to absorption (Hataet al., 1986).The concentration of essential fatty acids is reduced inoxidized oils, since these fatty acids are prone to oxidation(Watanabe et al., 1983). Some of the oxidation products of oils arehighly reactive and they may destroy labile vitamins and reduceavailability of some essential amino acids in vitro, notably,lysine, methionine, histidine and cystine, which are damaged byperoxides (Roubal and Tappel, 1966; El-Lakany and March, 1974;Gardner, 1979).Other symptoms of toxicity of auto-oxidized lipid in fish42include reduced growth, increased mortality and for carp, musculardystrophy or "sekoke" disease (Hata and Kaneda, 1980; Hung et al.,1981; Cowey et al., 1984; Ketola et al., 1989). However, whenantioxidants (vitamin E or ethoxyquin) have been added to the dietno toxic symptoms have been observed (Hung et al., 1981; Cowey etal., 1984).Thus, the proper preservation of fish is very important.Spoilage of raw fish as a food source is undesirable, because thenutritive value is greatly reduced and the end products producedduring spoilage could result in toxicity and death. Temperature of processingThe basis for preservation of fish by drying is the removal ofsufficient quantities of water to lower the water activity (Aw) toa level which inhibits bacterial growth and slows chemical andenzymatic reactions. Some form of heat is usually involved, rangingfrom freeze-drying (-40°C with plate temperature of 20 to 40°C) tohot air drying at 150°C (Bligh et al., 1988).The effect of processing on protein quality has been discussedpreviously by Bjarnason and Carpenter (1970), Bender (1972),Priestly (1979), Tarr (1982), Sandfeld (1983), Opstvedt (1988),McCallum and Higgs (1989) and Phillips and Finley (1990).The heating of proteins causes denaturation, which is arupture of the secondary and higher structures. The temperature atwhich denaturation occurs varies for different fish proteins withinthe fish and within fish species, and it is related to theenvironmental temperature in which the fish live (Aitken and43Connell, 1979). Generally, about 90% of protein is denatured at 60to 65°C, while the remaining 10% (tropomyosin) may be held at 100°Cfor a prolonged period of time without being denatured (Opstvedt,1988).Denaturation of proteins in fish may occur over time duringfrozen storage as well. Many hypotheses have been proposed toexplain the bonding in aggregated-denatured proteins that occursduring frozen storage of fish (Powrie, 1973; Sikorski et al., 1976;Matsumoto, 1979; Shenouda, 1980). During freezing, aggregation-denaturation is mainly caused by the formation of hydrogen,hydrophobic and disulphide bonds which increase in number duringfrozen storage (Jiang et al., 1988).Freezing, particularly at slow rates, causes the formation oflarge ice crystals. This leads to the breakage of cells, rupture ofmembranes, release of enzymes and disorder of cells and tissue. Aswater moves out of the cells, aggregation of myosin, disruption ofhydrogen bonding, release of enzymes and an increase in soluteconcentration occurs. These factors all contribute to proteindenaturation (Sikorski et al., 1976; Matsumoto, 1979). Theliberation of lipid due to ice crystal formation may result in theformation of insoluble lipoprotein complexes. The increase insolute concentration within the cell may also enhance the rate oflipid oxidation. The oxidized products of lipids bind specificfunctional groups of proteins such as cysteine -SH (sulphydrylgroup), E-amino groups of lysine, the N-terminal amino groups ofaspartic acids, tyrosine, methionine and arginine (Shenouda, 1980).44Denaturation per se does not appear to affect proteinutilization by animals in a negative manner. For example, freeze-drying of fish which causes extensive denaturation of theactomyosin (Connell, 1962; King, 1970) was not found to reduceprotein quality compared with that of the raw fish protein (Coweyet al. 1971, 1974; Opstvedt et al. 1984; McCallum and Higgs 1989).Although, reductions in available lysine, in vitro digestibilityand overall digestibility of the whole diet have been reported whenfreeze-dried protein sources have been employed(Fosbol, 1985;Koizumi et al., 1990). Generally the latter effect has beenascribed to lipid oxidation in the stored freeze-dried material.The low moisture content of freeze-dried fish allows for increasedlipid oxidation and browning due to lipid and Maillard reactionsduring storage (Tarr and Gould, 1965; Gardner, 1979; Bligh et al.,1988). Non-enzymatic browning reactions, or Maillard reactions, aredeteriorative reactions which involve the carbonyl group ofreducing sugar and the amino group of an amino acid or proteinresulting in the reduced availability of amino acids (lysine) (El-Lakany and March, 1974). Thus, storage of freeze-dried fish meal,even at -20°C, without the addition of antioxidant is notrecommended due to nutritive losses resulting from protein-lipidand Maillard reactions (Tarr, 1982).Careful drying of fish between 60°C and 100°C has been shownto have little effect on the nutritive value of protein in fishmeal as assessed by growth and digestibility studies in poultry,mink, and salmonids (Sandfeld, 1983; Fosbol, 1985; McCallum and45Higgs, 1989; Pike et al., 1990).The heating of protein at 50°C results in increased hydrationand loss of crystalline structure. Between 70 and 80°C, disulfidesplitting and loss of tertiary structure occurs; at 80 and 90°C aloss of secondary structure disulfides occurs; at 90-100°Cintermolecular disulfides form; at 100-150° loss of lysine andserine and isopeptide formation occurs; at 150-250°C peptidizationand more isopeptide formation occurs and at 200-250°C pyrolysis ofall amino acids and residues occurs (Phillips and Finley, 1990).At temperatures > 100°C, protein suffers thermal degradation,particularly with the heating of fish meals with a low moisturecontent. Fish meals dried with excessive heat (104°C) in comparisonto low temperature processing (85°C) when fed to chickens haveresulted in reduced growth and availability of essential aminoacids (Clandinin, 1949). Similar results by Tarr et a/. (1954)confirmed the lower nutritive value of overheated herring meals.Poultry fed diets containing high temperature meal and excessivelydried low temperature meal (dried for 180 minutes) suffered fromreduced growth when compared to poultry fed the low temperaturemeal (dried for 60 min.). Bisset and Tarr (1954) reported that theavailability of essential amino acids, particularly tryptophan,methionine, lysine, threonine and valine, as determined bymicrobiological assays, was impaired in overheated meal (159°C)dried for 180 minutes as compared to meal dried for 30 and 60minutes at 159°C.Carpenter et al. (1962,1963) also reported that the46availability of tryptophan, arginine, methionine and lysine wasreduced in overheated (115-130°C) herring meal. They concluded thatthe reduced availability of amino acids resulted from theinteraction of carbonyl compounds which formed during lipidoxidation and degradation.Opstvedt et al. (1984) found a linear decrease in the contentof -SH (sulphydryl) groups and a concomitant increase in thecontent of S-S (disulphide) bonds when fish was heated from 50 to115°C. The number of disulphide bonds produced in the heated fishvaried between fish species. When the heated fish (95°C) wereincluded in the diet of rainbow trout, a reduction in proteindigestibility of 1.1% and amino acid digestibility of 0.1-2.4% wasobserved. A similar decrease in protein digestibility of 4.9% andamino acid digestibility of 2.0 to 11.1% was observed in rainbowtrout fed a diet containing drum-dried fish (steam at 145°C) ascompared with freeze-dried fish.March et al. (1985) found a decreased growth response (16-18%)in rainbow trout fed a diet with heat damaged fish meal (127°C)included as 30-45% of the dietary protein in contrast to trout fedthe control fish meal diet. Similar results were obtained withchinook salmon fed a diet with overheated fish meal (150°C) andthis treatment lead to a reduction of PER, NPU and growth comparedto the fish receiving the diet with low temperature (75°C) herringmeal (McCallum and Higgs, 1989). Reduced growth has also beenobserved in mink, poultry and salmonids fed diets with overheated(>100°C) fish meals (Sandfeld, 1983; Pike et al., 1990).472.2.3 Methods of Measuring Nutritive Value2.2.3.1 Digestibility and growth measurements (In vitro)The quality of the test protein may be determined with thetarget animal in question. In the case of salmonids, the initialfeeding response is stimulated by food movement, since they arevisual feeders. The pellets are taken into the mouth, masticated,and held in the mouth until the salmonid determines if they arepalatable or not. Highly palatable pellets are swallowedimmediately while non-palatable pellets are spit out. Thus, one ofthe initial steps is to determine the palatability of a feedingredient when incorporated at different levels in the feed i.e.determine if the salmonid will accept the food containing theingredient or not. If a feedstuff is unpalatable, then it may bepossible to use feeding stimulants to mask the distastefulsubstances (Jones, 1990).A feedstuff may appear from its chemical composition to be anexcellent source of nutrients but it will be of little actual valueunless it can be digested and absorbed in the target species. Henceanother step is to determine the digestibility of the feedstuff inthe salmonid (Cho et al., 1982, 1985; Hossain and Jauncey, 1989).The direct determination of digestibility involves measuringthe amount of a specific nutrient ingested and then subtractingthat which remains in the feces following digestion (NRC, 1983).The collection of the feces from the aquatic media in which fishreside without contamination of feces by uneaten feed has resultedin approaches other than those used to measure digestibility with48terrestrial animals and birds (Cho and Kaushik, 1990). The aquaticmedia also poses serious technical problems for the recovery ofurine and gill excretions.The direct method of measuring digestibility involves thecollection of all egesta and excretions. Elaborate metabolicchambers have been constructed that involve filtration of water aswell as confinement of an individual fish in a special chamber witha rubber diaphragm that allows for the separate collection offeces, urine and gill excretions (Smith, 1971; Ellis and Smith,1984; Cho et al., 1982, 1985). This method is very labour intensiveand the confined fish are often under considerable stress which hasbeen shown to affect gastric emptying time, reduce appetite(Talbot, 1985; Smith, 1989b), increase mucous secretion (Pickeringand Macey, 1977) and enhance the level of trypsin lost in the feces(Pedersen and Hjelmeland, 1988). Also force feeding of fish (commonpractice in metabolic chambers) has been shown to effectdigestibility (NRC, 1981; Ellis and Smith, 1984). Thus, confinementand force feeding of fish is not recommended for digestibilitystudies, because the increases in the endogenous materialassociated with fish stress (secretions in the intestinal tract,sloughed epithelial cells and debris, enzymes and back fluxes ofamino acids) will result in under-estimation of nutrientdigestibility.If no correction is applied to the endogenous materialcontained in the feces, the term apparent digestibility coefficientis used in contrast to true or corrected digestibility coefficient.49The determination of true digestibility values involves feeding thetest animal a protein-free diet to allow for the measurement ofendogenous secretions (Schneider and Flatt, 1975; NRC, 1981). Suchdiets have been fed to rainbow trout (Nose, 1967 as cited in NRC,1981; Watanabe and Pongmanerat, 1991), carp (Hossuin and Jauncey,1989), milkfish (Ferraris et al., 1984) and catfish (Wilson et al.,1981) with no problems. The protein-free diet may be fed toomnivores and herbivores but true carnivores that require a highlevel of dietary protein have not responded well to this type ofdiet. Hajen (1990) reported that chinook salmon in sea water fed anon-protein diet refused to feed and lost weight during theexperiment. Thus, due to the problems and intensive labour involvedin measuring endogenous losses and the high correlation betweenapparent and true digestibility, especially when diets of highprotein content are employed, most authors have reporteddigestibility coefficients based on apparent digestibility(Schneider and Flatt, 1975; Smith et al., 1980; NRC, 1981, 1983;Cho et a/., 1982, 1985).An indirect method of measuring digestibility which has beenapplied successfully to terrestrial animals has also been used withsalmonids (NRC, 1981). The method uses an indigestible indicator inthe diet which is concentrated in the feces as nutrients aredigested and absorbed. The total collection of feces is notrequired which reduces the amount of labour and expense involved indetermining digestibility. An effective indicator is indigestible,unabsorbable, remains homogeneously mixed with feed and digesta and50does not interfere with digestion and absorption of nutrients(Schneider and Flatt, 1975; Smith, 1989a). The apparentdigestibility coefficient of a diet can be calculated as:% Nutrient= 100-100 x x(%Indicator food) (%Nutrient feces)Digestibility (%Indicator feces) (%Nutrient food)Edin (1918, as cited in Cho et al., 1982) first proposed theuse of chromic sesquioxide (Cr 2O3) as an external indicator fordigestibility trials. Chromic oxide has proven to be an effectiveindicator for terrestrial animals (Stevenson and De Langen, 1960;Kolb and Luckey, 1972) and for salmonids (Inaba et al., 1962;Hastings, 1969; Austreng, 1978; Cho et al., 1982, 1985; Tacon andRodrigues, 1984; De Silva et al., 1990; Hajen, 1990).However, Bowen (1978) reported that the stomach contents oftilapia had a lower concentration of chromic oxide than the feed(6.35% chromic oxide) and he concluded that the fish probablyselectively rejected the chromic oxide by masticating the feed andrejecting part of the food. Foltz (1979), however, criticized Bowenfor using high levels of chromic oxide and mixing the diets asaqueous suspensions in a laboratory blender which could haveresulted in a non-homogeneous mix of food and chromic oxide.Further, he concluded that proper mixing, diet preparation(pelleting) and proper feeding (feed readily consumed by testanimal) are very important when using chromic oxide as a tool fordetermining digestibility. The high specific gravity of chromicoxide (5.21) in combination with its insolubility in water result51in the non-uniform mixing of feed and chromic oxide. In this study,chromic oxide recoveries from food were less than 20% when the foodwas spiked with chromic sesquioxide in an aqueous medium. Bycontrast, recoveries of chromic oxide from spiked feces ranged from91.4-106.3% and from spiked feed(mash) ranged from 95.7-104.1%(Appendix Table 1). Similar differences in recoveries of Cr 2O3 fromfeces and feed have been reported by Saha and Gilbreath (1991a) whoattributed these to interference of mixed minerals formed duringdigestion with the analyses. Saha and Gilbreath (1991b) recommendedtaking great care when mixing chromic oxide and feed and ifpossible pelleted diets should be used to improve chromiumrecovery. Care should also taken during the analysis of chromicoxide whether assaying by either colorimetric or atomic absorptionspectrophotometry procedures to reduce the variation in analyticalrecovery (Saha and Gilbreath, 1991a).Tacon and Rodrigues (1984) found that chromic oxide proved tobe the most reliable indicator in terms of reproducibility.However, if chromic oxide concentrations in the diet exceeded 2%they reported that nutrient digestibility was significantlyincreased. Also, studies using the total collection of feces haveresulted in 95-100% recovery of chromic oxide (Cho et a/., 1982,1985; De La Noue and Choubert, 1986; Hajen, 1990) these findingsdemonstrate that chromic oxide is an effective external indicatorfor fish digestibility studies.In both indirect and direct digestibility studies of fish, itis essential that the feces obtained should represent52quantitatively the undigested residue of the food consumed. This isdifficult to achieve because of the aquatic environment andconsequently, many methods have been proposed for fecal collection.The subject of fish fecal collection has been reviewed by Austreng(1978), Windell et al. (1978b); Cho and Slinger (1979); Cho et al.1982; Vens-Cappell (1985); Spyridakis et al. (1989); and Hajen(1990).Briefly, collection of feces directly from the fish by■abdominal pressure, intestinal dissection and anal suction (Inabaet al., 1962; Austreng, 1978; Windell et al., 1978b; Vens-Cappell,1985; Spyridakis et al., 1989; and Hajen, 1990) can result in underestimation of nutrient digestibility for several reasons.Contamination of feces with body slime, urine and sexual productsincrease the level of nutrients in the feces which are not ofdietary origin (Cho and Slinger, 1979). Stress of capture,anesthetics and abdominal pressure increase the amount of fecesthat are normally expelled so that incomplete nutrient absorptionmay occur (Austreng, 1978; Leid et al., 1982; Spyridakis et al.,1989). The manual dissection of feces from the lower intestine isa terminal process and the fecal matter collected in this mannermay also be contaminated by endogenous debris, enzymes, urine,mucous and/or excess nutrients that could have been absorbedfurther before expulsion of feces from the posterior intestine.Also, the small amount of sample collected and the possibility ofdiurnal variation in digestibility as observed in domestic animals(Saha and Gilbreath, 1991b) and fish (Inaba et al., 1962; De Silva53and Perera, 1984; De Silva et al., 1990), suggest that the samplecollected by intestinal dissection may not necessarily represent"normal" fecal material.To overcome the problems of stress, repeat sampling andincomplete nutrient absorption, feces have been collected post-defecation from the water by settling column (decantation) (Cho andSlinger, 1979; Cho et al., 1982, 1985; Hajen, 1990), and bycontinuous filtration (mechanical rotating screen) of the effluentwater (Choubert et al., 1979, 1982; Spyridakis et a/., 1989). Themajor drawback to both of these methods is that soluble compoundsmay be lost in the water (leaching), and this effect may lead toover-estimation of nutrient digestibility. Windell et al. (1978b)found that apparent digestibilities for dry matter, protein andlipid increased by 11.5, 10.0 and 3.7%, respectively, after fecesremained in the water for one hour. Little or no additionalleaching of nutrients from the feces into the water occurred after4 hours. Smith et a/. (1980) also reported significant leaching ofsolubles from feces into the water. Even after feces and water werecentrifuged for 30 minutes at 3500 G, 50% of the soluble nitrogenwas retained in the water (supernatant). In this regard, it shouldbe mentioned that Cho and Slinger (1979) and Cho et al. (1982)recommended that the collected feces and water be centrifuged forat least 20 minutes at 10,000 G. Thus, the low speed ofcentrifugation used by Smith et al. (1980) may account for the highlevel of nitrogen which remained in the supernatant.The major cause of nutrient leaching from naturally voided54feces is break-up of the feces by physical handling (Cho et al.,1982; Vens Cappell, 1985; Talbot, 1985). This can occur during thenetting of feces (Windell et al. 1978b) or as a consequence of slowor incomplete settling of feces into the collection columns(i.e.improper flow rates and rough pipe fittings) (Spyridakis et al.,1989; Hajen, 1990).Proper settling of feces (transport of intact fecal pellets)has resulted in digestibility estimates for crude protein, drymatter (organic matter) and lipid that are not significantlydifferent from digestibility values obtained by intestinaldissection and in some cases stripping (Cho and Slinger, 1979; Choet al., 1982, 1985; Talbot, 1985; Cho and Kaushik, 1990; Hajen,1990). Thus, the settling column system allows for the continuouscollection of intact feces, without the handling stress associatedwith other collection methods. Many authors have adapted the"Guelph" settling column for the collection of feces from variousfish species (Spyridakis et al., 1989; Hossain and Jauncey, 1989;Ng and Wee, 1989; Hajen, 1990; Anderson et al., 1991; Watanabe andPongmanerat, 1991).Most of the reliable digestibility values for common fish feedingredients have been determined by comparing the digestibilitiesof a reference diet and a test diet composed of 70% reference and30% test ingredient. This is because few potential, usefulingredients can be used as the sole source of dietary protein. Theuse of a reference diet assumes that there are no interactionsbetween the components of the diet during digestion (Cho et al.,551982, 1985; Hossain and Jauncey, 1989). As mentioned above, theusual mixture is 70% reference diet to 30% test ingredient,although, Spyridakis et al. (1988) and De Silva et al. (1990) havereported varying digestibility estimates within the level of testingredient was increased in the test diet. De Silva et al. (1990)recommended that the inclusion of the test ingredient be at the 15-20% level for more reliable digestibility estimates. However, thedifferences in digestibility estimates were obtained only when thetest ingredients were of plant origin (i.e. high carbohydratelevel). With respect to this, progressive increases in the dietarylevels of carbohydrate have been found to decrease overall dietdigestibility in fish (Rychly and Spanoff, 1979; Steffans, 1989).In contrast, Anderson et al. (1991) reported that the level ofsubstitution of the test ingredient had no significant effect ondigestibility values for 16 feedstuffs in talapia when each wasincorporated into the reference diet at levels ranging from 20% to60%. Similarly, results by Cho et al. (1982,1985) and Hajen (1990)for salmonids have shown that the apparent digestibilities of eachof the test ingredients are additive and when totalled they equalthe digestibility value of the whole diet. Thus, the inclusion ofa reference diet in digestibility studies, allows measurements ofnormal feed intake and growth rate and hence the nutritionaladequacy of the experimental diets can be evaluated.The net absorption of amino acids from the digestive tractinto the blood of fish can be monitored using cannulated fish.McLean and Ash (1990) developed a method to cannulate the hepatic56portal vein of rainbow trout while the fish were able to swim andeat freely. Problems with the method involved a high initialmortality rate, failure of some fish to initiate feeding andrupture of sutures in some fish which had to be killed.Growth refers to the ways by which an animal increases in sizeand it is commonly measured as weight gain of the whole animal.Biological methods for determining the nutritive value of test feedingredients in diets involve assessment of protein efficiencyratio, net protein ratio, biological value (apparent and true), netprotein utilization, specific growth rate, feed efficiency, andenergy conversion efficiency. All of the procedures have been usedin fish studies (McCallum, 1985; Parker, 1987).A test protein source may be included in the diet at variouslevels to rate the dietary protein on a relative percent scale(March et al., 1985). Growth experiments are usually long term andlabour intensive so studies have been conducted to develop rapidchemical in vitro methods to assess nutritive value (Bodwell,1977). Enzymatic digestibility methods (In vitro)Since processing and storage conditions can have adverseeffects on the nutritional quality of protein, especiallydigestibility, considerable effort has been made to develop invitro assays for protein quality. Animal assays are not suitablefor routine monitoring because of the time and expense involved.Therefore, more rapid in vitro assays have been developed forassaying protein digestibility that rely on the use of proteolytic57enzymes (Swaisgood and Catignani, 1991).One of the oldest tests for assessment of in vitrodigestibility is the pepsin test. The official AOAC (1990) methoduses 0.2% pepsin with activity 1:10,000. However, this method hasbeen criticized for not being sensitive enough and consequentlydilute pepsin methods have been developed and used for fish mealquality assessment (011ey and Pirie, 1966).Lovern (1965) reviewed the efficacy of the pepsin test inrelation to varying concentrations of pepsin. He found thatpepsin concentration of 0.2% was unable to distinguish differencesbetween low temperature fish meals and overheated (1 and 5 hours at120°C) fish meals. Alternatively, when he used a pepsinconcentration of .0002% he was able to detect differences inprotein quality since pepsin digestibility values were 93.1, 90.6,and 85.0%, respectively. Also the .0002% pepsin values correlatedwell with animal feeding trials, available lysine and availablemethionine values. However, he concluded that the pepsin methodmeasured the solubilization and not the digestibility of theprotein.The pepsin test has been shown to correlate well with othertests (microbiological and poultry assay) for nutritive value whenapplied to Atlantic herring meals. For example 0.2% pepsindigestibilities were found to range from 93.9 to 95.4% whereas0.002% pepsin digestibilities were noted to vary between 90.0 and91.7% (Power et a/., 1969). This was also found to be true forPacific herring meals (values ranged from 92.1 to 93.0% were58obtained with 0.2% pepsin) (March et al., 1963).Pepsin digestibility values (.0002% with 1:10000 pepsin) havealso been found to correlate highly with chick growth assays, PER,and NPU when the chicks have been fed a variety (20 samples-3 fishmeals) of diets containing animal by-product meals (Johnston andCoon, 1979).The pepsin digestibility test has been used in Denmark andNorway to assess fish meal quality and high quality meals havegreater than 94% digestibility (Smith, 1990). The method has alsobeen adopted by the U.S.A. and only fish meals that exceed 92.5%digestibility are being used in diets for salmonids (NRC, 1981;Hardy and Masumoto, 1990). It should be mentioned that care must beexercised when comparing pepsin digestibility values becauseincubation temperature, pepsin strength (concentration andactivity) and duration of incubation greatly influence thedigestibility values (March and Hickling, 1982). The method hasbeen criticized for lack of sensitivity when comparing good qualityand excellent quality fish meals. Although differences in fish mealquality have been detected by dissimilar performance of fish infeeding trials, pepsin digestibility tests have not been found byHardy and Masumoto (1990) to be useful in this regard.The pH-stat method proposed by Milhalyi (1978 cited byRothenbuhler and Kinsella, 1985) is another enzymatic digestibilitytest. The method was modified and adopted by Pedersen and Eggum(1983) to overcome the buffering capacity of some proteins (O'Hareet al., 1984) that resulted in underestimation of protein59digestibility when the multienzyme pH-drop method of Hsu et a/.(1977) was employed.The pH-stat method utilizes a multi-enzyme combination (i.e.porcine trypsin, peptidase and bovine alpha-chymotrypsin) to assessprotein quality. As the proteolytic enzymes hydrolyse the proteinand peptide bonds, hydrogen ions are liberated into solutioncausing a decrease in the pH of the solution. In response to thisdecline in pH, base is added to maintain a constant pH at 8.00 forthe 10 minute test (Mann, 1988).The pH-stat test has proven very useful in predicting thenutritional value of various protein sources (Pederson and Eggum,1983; Rothenbuhler and Kinsella, 1985; Ritter et al., 1987; Lathiaand Koch, 1989). It is currently being used as a standard qualitycontrol test by the Danish fish meal and oil industry (Sandfeld,1983).The pH-stat test has compared well (correlation r > 0.90) within vivo experiments using rats fed various plant and animal proteinsources (Pederson and Eggum, 1983; Mann, 1988; Eggum et a/., 1989).Moreover this test has been useful in predicting the potentialnutritive value of various quality fish meals. However, the pH-statmethod overestimated protein digestibility of fish meals producedfrom spoiled raw materials when compared to digestibility valuesobtained by rainbow trout using the same fish meals (Sandfeld,1983).Recently, an in vitro method to simulate digestion in therainbow trout gut was developed by Grabner (1985). The method60involves the use of digestive enzymes extracted from the gut ofrainbow trout or carp which allows the monitoring of proteindigestion into amino acids (Grabner and Hoffer, 1985). A similarprocedure using extracts from carp intestine has been used todetermine the digestibility of a wide range of protein sources. Theresults have been shown to compare closely to digestibilitymeasurements conducted in carp (i.e. r = 0.99) (Eid and Matty,1989). Similar high correlations have been reported whendigestibility values determined using rainbow trout pyloric cecaextracts (pH-stat method) were compared to digestibility valuesmeasured with rainbow trout for various fish meals (Haard 1991).The problem with all initial enzyme catalytic rate assays isthat careful attention must be paid to parameters such as proteinsubstrate concentrations, enzyme concentration, protein/enzymeratios, ratios of various activities for multi-enzyme assays, andreaction conditions such as pH, ionic strength, temperature andduration of assay (Swaisgood and Catignani, 1991). Thus chemical invitro methods provide initial information on the potentialnutritive value of a protein but it is often desirable andessential to explore the nutritive value of proteins in vivo.61Chapter 33.0 EXPERIMENT 1 - Effect of methodology on determination of totalvolatile nitrogen and trimethylamine levels for previouslyfrozen Pacific herring (Clupea haraengus pallasi) stored at 2-5°C for up to 15 days.3.1 INTRODUCTIONHerring are often frozen for prolonged periods when surpluscatches occur and processing equipment cannot keep pace with supply(Connell, 1990). Fish by-products (surplus catches, unsold frozenstores, intact males and roe-removed females) are often directed toreduction plants for conversion into fish meal, which is then usedprimarily as a protein source in animal feeds. The fishing ofherring specifically for use directly in fish meal production isprohibited both in Norway and in Canada (Halland et al., 1988;1989;DPA, 1988).It is important to assess the quality changes whichoccur in fish before freezing, during frozen storage and afterthawing. Few studies have examined the spoilage of previouslyfrozen fish after defrosting and storage at elevated temperatures(Shewan, 1961; Gould and Peters, 1971; Flick et al., 1986; Hallandet al., 1990).In relation to the chemical methods which have been employedto assess fish spoilage, total volatile basic nitrogen (TVN,TVBN)and trimethylamine (TMA) measurements are two of the most commonlyused. Both methods have been criticized for lack of sensitivityduring the earlier stages of fish freshness but both parameters aregood indicators of advanced fish spoilage (Reay and Shewan, 1949;62Farber, 1965; Shewan et al., 1971; Flick et al., 1986; Connell,1975; 1990).TVN is primarily composed of TMA and ammonia, but othersubstances may be present such as dimethylamine, methylamine, andformaldehyde (Reay and Shewan, 1949; Halland and Njaa, 1988),especially during frozen storage of gadoid fish (Castell et al.,1970; 1971; 1973). Norwegian studies with low temperature (LT)produced fish meals have indicated that raw material freshness isthe most important factor in producing high quality meals. Rawmaterial with TVN levels of ideally less than 50 but up to 90 mgper 100 g fish is the most desirable for the production of goodquality fish meal (Pike et al., 1990).TMA is found in very low levels in fresh marine fish but thisvolatile amine can accumulate to high levels during spoilage,mainly as a consequence of the bacterial (Alteromonus sp.)reduction of trimethylamine oxide (TMAO) by the enzyme triamineoxidase (Beatty and Gibbons, 1937; Tarr, 1939; 1940; Hedard et a/.,1982; Hobbs, 1987). The reduction of TMAO to TMA can also occur, toa lesser extent, by other chemical pathways (endogenous fishenzymes, heating, metal ions), but during fish spoilage thereduction resulting from bacteria is the major pathway (Castell,1970; Hebard et al., 1982). Bacteria can also produce TMA fromcholine and betaine (Dyer and Wood, 1947; Tokunaga et a/., 1977).The level of TMA formed during spoilage of herring on ice isgenerally not great enough to be analytically useful for monitoringfreshness (Hughes, 1959; Connell, 1975; 1990; Damoglou, 1980)63unless specific conditions occur. Larger amounts of TMA are knownto be formed in herring during anaerobic storage (i.e. polyethylenepackaging, storage in chilled sea water) (Smith et al., 1980;Fernandez-Salguero and Mackie, 1987; Huss, 1988). The recommendedupper limit for TMA content in good quality cold-water fishdestined for human consumption is 15 mg TMA-N/100 g fish (Connell1975).Various methods have been proposed for determining TVN and TMAlevels in fresh herring and other fish species. Such methods asdirect distillation (Lucke and Geidel, 1935; Haaland and Njaa,1988), steam distillation (Hjorth-Hansen and Bakken, 1947; Malleand Poumeyrol, 1989), microdiffusion (Beatty and Gibbons, 1937;Aksnes, 1989), colorimetric assay (i.e. picrate method, Autoanalyzer) (Dyer, 1945; Shewan et al., 1971; Ruiter and Weseman,1976), enzymatic (Wong and Gill, 1987; Wong et al., 1988),electrodes (Chang et al., 1976; Bagnasco, 1985; Ohashi et al.,1991), gas chromatography(GC) (Hughes, 1958;1959; Fernandez-Salguero and Mackie, 1987; Perez-Martin et al., 1987), highperformance liquid chromatography (Gill and Thompson, 1984) and GCheadspace analysis (Krzymien and Elias, 1990; Fiddler et al., 1991)have been used. The GC technique appears to be the most sensitiveand specific, although it is very time-consuming and requiresexpensive materials and apparatus (Malle and Tao, 1987). Also,difficulties with loss of sample response, tailing peaks andghosting phenomena have been reported during GC analysis ofvolatile amines (Keay and Hardy, 1972; Fiddler et al., 1991).64Caution should be used when comparing TVN and TMA valuesbecause these values vary markedly, depending on the method used(Botta et al. 1984; Connell 1990; Clancy et al. 1990). Thus, thefollowing study was undertaken to examine the merits of variousmethods for determining the levels of TVN and TMA during thespoilage of previously frozen Pacific herring destined for fishmeal production.3.2 MATERIALS AND METHODS3.2.1 Raw MaterialsPacific herring were caught by commercial fishermen using gillnets on March 28, off the central section of the Vancouver coast.The fish were stored in refrigerated sea water until they weresexed by machine. Thereafter the ripe males were bagged(polyethylene) and snap frozen at -40 °C for 24 hours. The fishwere then stored at -20°C until processed. For the experimentalwork, 50 fish were selected at random from the common lot, placedin a polyethylene bag, sealed and stored at 2-5°C for the durationof the experiment. The spoilage of the herring over a 15-day periodwas determined by monitoring the levels of TVN and TMA using sevendifferent methods.3.2.2 Chemical AnalysesOn day 0 and 1, and between day 4 and 15, 3 fish were selectedat random from the 50 fish and passed twice through a meat grinderwith 4.8 mm diameter holes. The sample was mixed thoroughly,wrapped in an oxygen impermeable film (aluminum foil), refrozen,stored at -40 °C and analyzed as soon as possible. TVN and TMA65levels were determined on 10-50 g of minced fish using thefollowing methods: a) TVN by direct distillation with MgO using aKjeldahl distillation apparatus and titration as described byWoyewoda et al. (1986). The analyses were conducted in triplicatewith the aid of pH meter and 0.01 ml graduated burette with a microtip to improve precision.b) TVN and TMA by microdiffusion with K2CO3 . This involvedtrichloroacetic acid (TCA) extraction followed by microdiffusion ina Conway chamber (Conway and Byrne, 1933 as modified by Obrink,1955) and titration as described by Beatty and Gibbons (1937) asmodified by Murray and Gibson (1972a) and Malle and Tao (1987).Triplicate analyses were performed and the microdiffusion unitswere cleaned as described by Conway (1957). Sample prepartionentailed homogenization of 50 gm of fish (Polytron-60 sec.) in 150ml 7.5% trichloroacetic acid solution. The homogenate was thenfiltered through Whatman #4 filter paper and stored in glass jarsat 4°C. For TMA analysis 0.5 ml of neutralized 40% formaldehyde wasadded to the outer chamber of the dish to block the release ofammonia, primary and secondary amines (Conway and Byrne, 1933;Benoit and Norris 1942) prior to incubation at 37 + 1.0 °C for 2hours.(c) TVN and TMA by steam distillation with NaOH. The methodproposed by the Codex Alimentarius Committee in 1968 as modified byBillon et al. (1979), Vyncke et al. (1987) and Malle and Poumeyrol(1989) was used. Briefly, sample preparation involvedhomogenization of 50 g of fish in 150 ml 7.5% TCA solution66(Polytron-60 sec.) which was then filtered and stored at 4°C.Steam entrainment was performed until a final volume of 50 mlwas collected (12 minutes total) using a Kjeldahl-type distillationapparatus (Buchi 325) with a distillate flow rate of 10 ml/min. asrecommended by Antonacopoulos (1985) and Antonacopoulos and Vyncke(1989). To determine the TMA content, 20 ml of formaldehyde isadded to the 25 ml of filtrate in the distillation tube asrecommended by Malle and Tao (1987).3.2.3 Statistical analysesThe TVN and TMA results were subjected to One-way analysis ofvariance (ANOVA) with method as the main effect. When ANOVAindicated a significant difference among the methods, Bonferroni'stest with P=0.05 was used to detect significant differences betweenthe treatment means. Correlation coefficients were determinedbetween the TVN and TMA method values over the storage time.3.3 RESULTS AND DISCUSSION3.3.1 Total Volatile NitrogenThe TVN level increased during storage as monitored by allmethods, especially after 8-9 days (Fig. 4). The sudden increase inTVN presumably resulted from an increase in bacterial numbers,since a good correlation between TVN level and bacterial numbershas been reported previously (Reay and Shewan, 1949; Lucke andGiedel, 1935; Hjorth-Hansen and Bakken, 1947).The trend of increase was very similar regardless of themethod of analysis (Fig. 4). The actual values for TVN after 15days of storage, ranged from 80.2 to 113.6 mg N/100 g (as is67'CS cf)" c,c?	a) 	 =	E 	 czt 7:8 "°	co — 	 = Lrl— Z cucmD > E .2	 z 	 E4	L 	 4ae g"	0 	 (i)	CD 4-, 	 = ;—■ ;—,CI) 61) cLU 420 4-1L). E = 0Tr 	 a) 1) 'CI> N•41 0al cl) c+.4-6 >.-c)> -0 cuc., uzi ccz$ 0 -4—,0•= `14 ,..‘= E FI- •,-as 0 ;-0 	 1) -1..‘(1) CD68basis), depending upon the method employed. The overall pattern forthe results was very similar irrespective of the method employed,except for the method using direct distillation with MgO (Table 3).The high TVN values observed with direct distillation with MgOmay be erroneous because protein breakdown during atmosphericdistillation with MgO has been reported previously (Pearson andMuslemuddin, 1968; 1969). Lucke and Geidel (1935), for instance,found that as distillation time increased with MgO, some ammoniawas produced due to the breakdown of protein in the intact fishsample used. Further, Botta et al. (1984) reported an increase of10.9 and 14.7% in TVN values when using steam distillation with MgOand direct distillation with MgO after protein (bovine serumalbumin) was added during analysis. TCA extraction plus microdiffusion with K2CO3 were observed to result in the lowest increasein TVN value of 3.1%.The coefficient of variation (CV) for TVN determination wasgreater for the method involving steam distillation with NaOH thanfor all the other methods (Table 3). Malle and Poumeyrol (1989)reported a CV for TVN of .5625 (mean of three fish species) whenusing the same method, which is slightly lower than the CV valuefor herring found in the present study. The strong alkalizingability of NaOH, as compared to MgO, combined with a higher pHduring distillation results in the production of additional ammoniafrom secondary deamination of nitrogenous compounds. This effectmay explain the high variation in results when using NaOH(Antonacopoulos and Vyncke, 1989). Thus, Antonacopoulos and Vyncke—4 0ItS 0W r-iza)4-1 ZE-4O• 0z69	CO 	 coC•1 	 'o	Ul 	 1/40 	 In 	 1C)• • 	 •	 •0 CI CV 1•1CV 	 o CO• •	 •	 •In 	 •:11 	.44	N .0 	 .0 	 .0tc,Nin Li) CI CV• • 	 •CO CV CI I-IIn 	 •:1' 	 •cr 	 •ZreCV 	 CN 	 ••I▪ •:1' 	 CIOZ▪ 0I 	0r-i• 	I 	 ••••I1-10• 0 0• 4-14.4 	 •1•1U)• 	U)•••1 	 0 	 0'0 	 4-4 	 4-1 	 4-)4-i 	 4•4 	 -r14.) 	 "0a) 0 0• 0')-1i 	 U 	/4• CD4.)O 0 W4.-▪ ) 4-4 4) 4)>4 	 >4 	 >43 	 a) 	 a) 	 a)O 4c4 	 4c4tr U U C.)E-4 	 E-4 	 E-•C.) 0 CICO 	 l"••• 	 CO• • 	 •cn 'o 0CI .44 0• • 	 •CV CV CNoO rnN 	 sZI'r•• 	 N 	 •:1*(1/410 	 1-1 %.0t'0z0:1c0I 	 I	 4.1ttO• 0Cl)	 U) 	 •Pi4-1 	0 4-)4-1 	 U)4-1 	 4-1'0 '0O 0$.4a)4.1O 0 CO4-▪ 	) 	4-)>4 	 >4 	 >4a) 	 a) 	 a)fr4 	 0:4C.) 	 C.) 	 C.)E-i 	 EIn0OVI44r.a)4-)a)4-44-4• -I'04-)• W• 0•r-I4-4-4-4 	 0• •,-)01 •4-),-I 4-)U) 	 •••Ia)a) 4-) a)• •c:$• N• -el $.4a) 	 414)4.4 04.-) $4a) )4 01--1 CD 4-)4-1 wa)O 0azs ralU)	 0.CU 4.)as4_) b •4•4cuis• 0cU O 04 4-3U)	 4-)•cl4-) 	 a)O Ea -H• 0 0U) $-4 4-4• a) 4-I• 4 cua) -1-3z—00r4zU)U)IUcU4-40'ri04-)a)70(1989) proposed a detailed working procedure for TVN determinationusing controlled steam distillation with MgO to help reduce thevariation between replications, operators and laboratories.The TVN results indicated that the differences among themethods were relatively constant over the range of measurements(correlation coefficients of .994-.985). The results indicate thatTVN is not a good indicator of initial spoilage but is a goodindicator of advanced spoilage. This finding agrees with those ofearlier studies (Beatty and Gibons, 1937; Reay and Shewan, 1949).In general, the levels of TVN for Pacific herring stored at 2-5°C followed a similar trend. However, they were slightly higherthan those reported for herring stored on ice (Hjorth-Hansen andBakken, 1947) and for herring stored anaerobically at 2°C (Hallandand Njaa, 1988).3.3.2 TrimethylamineThe levels of TMA paralleled the levels of TVN during storageof herring at 2-5°C (Fig. 4). The initial values for TMA were lowerthan those for TVN and only after 8 days was a significant amountof TMA present. The trends for TMA levels remained the sameregardless of the method of analysis (Fig. 4). These resultsindicated that the differences between the methods were reasonablyconstant (correlation coefficients of .890-.979).The overall values for TMA did not vary markedly between themethods, with the exception of microdiffusion with KOH, whichresulted in slightly higher values than those obtained with theother methods. However, the values were not significantly higher71than the others (Table 3). This finding contrasts with those ofprevious authors who have found that KOH is a superior releasingagent than K2CO3 . If other primary and secondary amines are presentduring analysis, they may interfere with K2CO3 , thereby resulting inelevated TMA values (Hashimoto and Okaichi, 1957; Tozawa et al.,1971). DMA is produced in appreciate amounts during frozen storage(-6 to -29°C) of gadoid fish (cod, hake, pollock) by the enzymaticreduction of TMAO (Tozawa et a/., 1971; Mackie and Thomson, 1974).However, other species produce no DMA during frozen storage,especially at temperatures below -29°C (Smith et al., 1970; Hebardet al., 1982). In this experiment interference from DMA should nothave been a problem, since it has been shown by Reay and Shewan(1949) and by Fernandez-Salguero and Mackie (1987) that duringspoilage of herring DMA is produced only in small amounts (< 1.0 mgN/100 g).It is not known whether the difference between the TMA valuesobtained with micro diffusion using the reactive agents KOH andK2CO3 in the present study represent DMA production. Thesimultaneous determination of TMA and DMA in fish is possible whenusing KOH and K2CO3 with the picrate colorimetric procedure (Castellet al., 1974; Woyewoda et al., 1986). Cobb et a/. (1973) havedeveloped a technique for the measurement of TMA and DMA usingmicro diffusion with K 2CO3 and Na3PO4. They found that microdiffusion with K2CO3 resulted in a 67.5 % recovery of DMA in thepresence of formaldehyde (FA). The picrate procedure was notemployed in the present study because of the explosive nature and72health risks associated with picric acid. Spiking studies withvarying ratios of DMA and TMA would have to be conducted with thereactive agents KOH and K 2CO3 to establish a method of determiningDMA with micro diffusion. For the specific, quantitative analysisof DMA levels in fish GC is recommended (Keay and Hardy, 1972;Perez-Martin et al., 1987; Fiddler et al., 1991).The concentration of KOH with respect to "TMA extractability"from fish muscle has been discussed in the past (Shaw et al.,1977). Murray and Gibson (1972 a;b) found that higher readings(picrate method) of TMA were obtained if the concentration of KOHwas increased from 25 to 45%. Castell et a/. (1974) studied theeffects of varying concentration of alkali solutions (KOH and K 2CO3 )on the optical density of TMA and DMA picrates. They found that therelationship of color intensity of the picrate to K 2CO3 was linearbut concentrations greater than 30% KOH resulted in a non-linearrelationship especially if DMA was present in larger quantities.Keay and Hardy (1972) also reported that KOH resulted in higherrecoveries of TMA than K 2CO3 when using GLC. Bullard and Collins(1980) observed increased "TMA extractability" when using higherconcentrations of KOH and recommended the use of 45% KOH at -15°C(picrate method) for the extration of TMA when DMA is present inhigher quantities. Thus, the high concentration of the KOH solution(45%) used in the present study may explain the higher TMA valuesobtained by microdiffusion with KOH than by microdiffusion withK2CO3 •The microdiffusion method has been rejected by Spinelli (1964)73for the routine determination of TMA in fish, even though goodcorrelations have been obtained between TMA results based uponmicrodiffusion, steam distillation, picrate method and GC methods(Beatty and Gibbons, 1937; Dyer, 1945; Murray and Gibson, 1972a,b;Malle and Tao, 1987). Shewan et al. (1971) emphasized the need forproper cleaning of the apparatus to ensure accuracy of the data.Also they found that the microdiffusion method was simple,inexpensive, fairly accurate, reproducible and was used widely .With further modifications to the design (Obrink, 1955) andmaterial (polished plastic), the microdiffusion cells are lesslikely to become contaminated and they require only rinsing withwater and weak acid with indicator to ensure that the cells areclean. The use of micro magnetic stir bars, plastic pipettes,burettes and pH probes also results in accurate and reproduciblefindings. The inexpensive microdiffusion units (Botta et al., 1984)enable the precise and accurate simultaneous determination of TVNand TMA in a reasonable time period (2 hours at 37°C).The determination of TMA in Pacific herring using steamdistillation with NaOH resulted in the lowest values and thehighest C.V.. In general, the C.V. values for TMA determinationwere much higher than those for TVN (Table 3). Malle and Poumeyrol(1989) observed similar high C.V. values for TMA when compared tothe C.V. values for TVN when using steam distillation with NaOH.A possible reason for such high C.V. values for TMA is the useof large volumes of formaldehyde (FA) for analyses and the use of FAin the presence of monomethylamine and DMA have resulted in74anomalous readings of TMA (erratic results) (Benoit and Norris,1942). Gibson (1982) emphasized the importance of using the correctconcentration of FA in relation to the concentration of ammonia,DMA and TMA in order to obtain a constant recovery of TMA. Thus thedetermination of TMA in the presence of ammonia, mono and DMA is amore complex reaction than the determination of TVN.In the present study the lag period for TMA production waslonger (9 days) than the lag period of 5 days reported byFernandez-Salguero and Mackie (1987) for whole herring stored at5°C. The lower temperatures for the Pacific herring stored at 2-5°Ccould account for the lower TMA levels, since the production of TMAin herring is temperature dependent (Sigurdson 1947). The TMAlevels in the present study are similar to those noted for wholeherring stored on ice (Damoglou, 1980; Fernandez-Salquero andMackie, 1987) and in refrigerated sea water (Smith et al., 1980).Freezing fish (below - 6°C) reduces the bacterial populationby 60-90% but upon thawing and storage on ice, bacterialpopulations (Pseudomonas-Alteromonas and Moraxella-Acinetobactergroups) recover quickly and spoilage progresses at a similar rateto that observed for fish on ice that have not been frozen (Shewan,1962; Hobbs and Hodgkiss, 1982; Hobbs, 1987).The lag phase for bacterial growth in defrosted cod has beenfound to be increased by several days (2-5) after they have beenfrozen for 3 to 6 months. Experiments conducted at the TorryResearch Station on the spoilage behaviour of thawed cod on iceconfirm these findings and they show that freezing may increase the75lag phase for TVN and TMA levels as well (Shewan, 1961). Luijpen(1958) found that TMA and ammonia formation were suppressed for upto 5 days in previously frozen cod relative to non-frozen codstored at 2°C.The TMA results indicate that TMA is not a good indicator ofearly quality changes but it is a good indicator of advancedspoilage. This finding supports those of earlier studies(Sigurdsson, 1947; Reay and Shewan, 1949; Hughes, 1959; Smith etal., 1980).3.4 CONCLUSIONSAlthough the method of analysis definitely affected the TVNand TMA values obtained, all seven methods resulted in similartrends and each appears to be suitable for use as an index ofadvanced spoilage, provided that definite spoilage specificationsare established for each method.Both steam distillation and microdiffusion allow for thesimultaneous determination of TMA and TVN. Steam distillation withNaOH is a more rapid method (12 min.) but the results for TMA andTVN have a high C.V.. The less expensive microdiffusion with K 2CO3method results in similar TMA and TVN levels to those obtained withthe steam distillation with NaOH procedure but the C.V. values arelower in each case. Also, the microdiffusion method can beconducted in about 2 hours.The literature contains a wide variety of methods that havebeen applied to measuring quality changes in chilled fish. However,it is evident that due to several complicating factors (seasonal76changes, geographical location, size, rate of freezing, storagetemperature, pre-freezing history) it has proven difficult to finda suitable method or methods for routine commercial or evenresearch use. With respect to fish destined for human consumption,sensory analysis is still favoured. However, fish used for animaldiets may be in advanced stages of spoilage. Thus, it has beennecessary to investigate the efficacy of various chemical methodswith respect to their suitability for assessing advanced spoilageof fish.The levels of biogenic polyamines (putrescine, cadaverine,histamine) are also good indicators of spoilage for herring storedat various temperatures (Fernandez-Salguero and Mackie 1987; Aksnes1989). Accordingly, the levels of biogenic polyamines were alsodetermined in this study in previously frozen Pacific herringsubjected to various conditions of spoilage and subjected todifferent thermal processing regimes.77Chapter 44.0 EXPERIMENT 2. Influence of storage and processing conditionson the levels of total volatile nitrogen, trimethylamine,putrescine, cadaverine and histamine for previously frozenPacific herring (Clupea harengus pallasi).4.1 INTRODUCTIONHerring (by-product of the roe industry) are often frozenbefore processing into fish meal which is then primarily used as aprotein source in animal feed (Haaland et al., 1990). Almost allfish meal is produced by cooking the fish, pressing the cooked massto remove most of the oil and water, and drying the resultantpresscake to which a concentrated fraction of the aqueous pressliquor is normally added. The final product is ground, stabilizedwith antioxidant and bagged for storage (Fig. 2). The resultantfish meal, which is a brown powder, normally contains a high levelof protein and appreciable quantities of lipid and minerals(Windsor and Barlow, 1981; Barlow and Windsor, 1984; FAO, 1986).From the time of catching, fish undergo chemical changes dueto oxidation, the action of bacteria present on the surface and inthe gut of the fish and the action of enzymes within the fish(autolysis) (Pike et al., 1990). As the fish deteriorate, proteinis broken down to peptides, amines and free amino acids whichprovide nutrients for the pathogenic bacteria. Amino aciddecarboxylation by bacteria is the most common mode of synthesis ofthe biogenic polyamines (putrescine, cadaverine, histamine) asdiscussed previously.Trimethylamine oxide (TMAO) which is present in high78concentrations in marine fish, is also broken down into volatileamines. These consist mainly of trimethylamine (TMA) butdimethylamine (DMA), methylamine and ammonia are also produced(Hebard et al., 1982). The major factors leading to the formationof volatile and non-volatile amines are levels of free amino acids,bacteria, fish species and storage time and temperature (Klausenand Lund, 1986). Hence, the levels of volatile amines and non-volatile polyamines are useful as quality indices for assessingdecomposition of fish such as herring (Hughes, 1958; Ritchie andMackie, 1980; Fernadez-Salguero and Mackie, 1987).Storage of fish at temperatures greater than chill storage(0-5°) results in the rapid formation of non-volatile amines(Ritche and Mackie, 1980; Yamanaka et al., 1986; Middlebrooks etal., 1988; Hollingworth et al., 1990; Wendakoon et al., 1990) andvolatile amines (Hebard et al., 1982; Fernandez-Salguero andMackie, 1987; Veciana-Nogoues et al., 1990). Frozen storage at< - 6°C prevents the formation of both non-volatile and volatileamines (Mackie and Thompson, 1974; Kramer et al., 1977; Hebard etal., 1982; Baranowski and Pan, 1985; Baranowski et al., 1990) withthe exception of DMA (volatile amine) formation which occurs ingadoid fish stored at temperatures above - 30°C but below 0°C(Castell et al., 1971; 1973; 1974).The increased lag phase in the production of TVN and TMA(Luijpen, 1958; Shewan, 1961) and histamine (Baranowski and Pan,1985; Baranowski et al., 1990) due to freezing fish before chillstorage (0-5°C) is a consequence of the sensitivity to freezing of79the photobacteria responsible for the deterioration (Van Spreekens,1987).Few investigators have studied the formation of volatile andnon-volatile amines in thawed fish (Gill and Thompson, 1984; Farnand Sims, 1987).High levels of histamine have been found in fish meals whichhave been produced from spoiled raw materials, especially when thefish solubles are added back into the meal (Toyama et al., 1981;1985). Generally, an increase in processing temperature and/orduration of thermal processing result in depression of thehistamine level in the meal. Similarly, the higher the pH and/orlower the water activity, the greater the loss of histamineassociated with processing due to the non-enzymatic browningreaction (Maillard reaction) as well as from histamine binding withproteins (Toyama et al., 1982; Tanaka et al., 1986).Putrescine and cadaverine levels in fish are also decreasedduring thermal processing (Farn and Simms, 1987) However, thermalprocessing increases the levels of TMA and DMA in fish since TMAOis heat sensitive (Hughes, 1958; 1959; Hebard et al., 1982;Gallardo et al., 1990).The methods available for the analysis of biogenic aminesrange from bioassays to chemical assays that include fluorometric,enzymatic and chromatographic procedures (Taylor, 1986; Taylor andSumner, 1987).Fluorometric procedures are the most widely used for thedetection of histamine. All are based on the condensation of80histamine with o-phthalaldehyde to yield a fluorophore (Taylor andLieber, 1977). The methods differ only in the procedure forcleaning the sample to remove substances that might otherwiseinterfere with the condensation reaction. The official AOAC methodrelies on an anion exchange resin based upon the method ofStaruszkiewicz et al. (1977). However, interferences have beennoted with histidine, histidyl peptides and other amines includingspermine and spermidine (Taylor, 1986). Lerke et al. (1983)developed a rapid and sensitive enzymatic method for detectinghistamine in raw and heat-processed fish.Chromatographic methods for histamine analysis in fish requireexpensive equipment but they allow the quantitative detection ofnot only histamine but also putrescine, cadaverine, trimethylamine,dimethylamine, tryptamine, spermidine, spermine, tyramine andagmatine. Gas chromatographic (GC and GLC) procedures require thatthe amines be converted into volatile derivatives for detection(Mita et al., 1980; Yamamato et a/., 1980; 1982; Staruszkiewicz andBond, 1981; Yamanaka et al., 1989; Wendakoon et al., 1990). Highpressure liquid chromatography (HPLC) requires the derivatizationof the amines into products that can be detectedspectrophotometrically or fluorometrically (Mietz and Karmas, 1977;Karmas and Mietz, 1978; Mietz and Karmas, 1978; Hui and Taylor,1983; Gill and Thompson, 1984; Rosier and Peteghem, 1988;Hollingworth et al., 1990; Yen and Hsich, 1991).This study was undertaken to further evaluate the extent ofspoilage of thawed Pacific herring during chill storage by81monitoring the formation of both volatile and non-volatile amines.Also, the levels of amines in the herring meal after thermalprocessing were determined in an attempt to better chemicallydefine fish meal quality in relation to raw material freshness.4.2 MATERIALS AND METHODSThe levels of putrescine, cadaverine, histamine, totalvolatile nitrogen and trimethylamine were determined in previouslyfrozen Pacific herring (Clupea harengus pallasi) stored at 2-5°Cfor 0, 8 and 12 days before and after thermal processing at bothlow and high temperatures.4.2.1 Raw Material and Processing ConditionsThe source of Pacific herring and methods used for storage andprocessing have been described previously (section 3.2.1). Theexperimental design included three degrees of freshness and twoprocessing temperatures. In this regard, the herring were thawedand then stored for 0, 8 or 12 days at 2-5°C, with each treatmentprocessed at two temperatures, low temperature and high temperatureand a freeze-dried batch provided the seventh treatment (Table 4).The herring meals were prepared from the common lot of rawfish in a continuous pilot scale fish meal manufacturing machine.The machine consists of a steam-jacketed cooker, screw press, andsteam-jacketed rotary dryer (Claggett, 1968). Cooking and dryingtemperatures were both set at either 75°C or 100°C for each batchof fish. The press liquor was not added back to the press cake mealbut the screened solids(Kason Vibroscreen with mesh U.S. #10-2.0mm) from the press liquor were added to the dryer (Fig. 2). The821-- 1"" co 	 coIPCM 	 CVT 0      o	 to 	 0 	 to 	 0 	 0o t•-. 	 0 	 N. 	 0 	 .1-,- 	 ,- 	 ,- 	 1v 	 v 	 -c3 	 V 	 -c3 	 -Do 	 o 	 a)a) 	 a) a .— a wL... 	 T.: C 	 1.- C 	 .17-L. 	 1-3V 	 V 	 V .17 	 1.- V 	 V.0a) 	 2 m iii 	 2 a) 	 2,21— .0 	 .c 	 2 	co= 	 = 	 cm = cm= 	 = 	 4-7.1 	 7E; a 'co -ca. c a 4) _▪ c te -:-.. lii .L c 	 L.. co	 o	 s_ 	 Vc&-. c• •— 	a)• 	 a) cl) 	 a) a) 	 8 ,Lz. 	 CD " 	 CD 0)CCa.. t 	 gL. i. 	 Ca. ift 	 Ca. L-4- _ 	 L... ._E w E a) E 	 E 	 E 	 EL V ",..,a) -c 0 c a) E 2 E m -o a) v 	 a)}' = 	 = 	 a) 	 a) 	 a) .cN_C 	 _C 	 = r = G.,..c co 	..c 73	 3 o 	 a)0)3 ci3 	  13 a a,o 2 .°) 0 	o cp	 o o. .— o. 3••• as—I .«- 2 .41"-- JE 2— E _1 co 2 co LL L..a)o01--IILL.i--Jto1Cl)i-- JcoI 	 N.2 	 ICO 	 2I- 	 02 	 Li_83meal was held in the dryer for 15 to 20 minutes until the moisturecontent of the meals was approximately 10%. The meals were thencooled in a vertical cooler by ambient temperature air flow, andsubsequently each treatment was stabilized with ethoxyquin (0.02%), and finely ground (Fitzmill, model JT) to pass through a U.S.#20 screen (0.841 mm). Fish for the freeze-dried protein source(FDH) were thawed, ground in a meat grinder (Butch Boy model TCA32with 12.5 mm die) refrozen on trays at -20°C for 24 hours andfreeze-dried at -40°C with a plate temperature of 23°C. The FDHwas then ground (Fitzmill) to pass through a U.S. #20 screen andstabilized with ethoxyquin (0.02 %). All protein sources were thenvacuum packed and stored at -40°C until further analysis.4.2.2 Chemical AnalysesPutrescine and cadaverine determinationPutrescine (1,4-diaminobutane) and cadaverine (1,5-diaminopentane) were extracted from the herring (10.0 g) andherring meal (3.0 g) using methanol (50 ml-ACS grade FisherScientific). Samples were homogenized (Polytron-2 min.), made up to100 ml with methanol, filtered (#4 Whatman) and stored in amberglass bottles at 4°C for a maximum of 1 week.An internal standard (0.5 ml hexanediamine working solution 10ug/ml prepared from a stock solution of 100 mg hexanediamine madeup to 100 ml with 0.1 N HC1) was added to 10.0 ml of sample extractalong with 0.5 ml 1.0 N HC1. The solution was evaporated to drynesson a Rotovap (50°C - 50 min.). The derivatization step involved theaddition of 1.0 ml ethyl acetate and 300 ul pentafluropropionic84anhydride (99 % - Sigma) to the dried extract which was furtherheated at 50°C for 30 min. in a water bath. The volume was reducedin a Rotovap at 50°C to remove excess solvent and reagent. Theresidue was dissolved in 1.0 ml 30% ethyl acetate in toluene.Sample cleanup involved passing the reaction mixture throughan aluminum oxide column using a solvent combination of hexane (10ml), extract and increasing volumes of 30% ethyl acetate in toluene(3, 9 and 15 ml). The effluent from the column was collected andstored in amber glass bottles at 4°C.Samples (1-2 ul) were injected into the injector port (200°C)of the chromatograph (Shimadzu -GC14A) equipped with an OV-225fused silica capillary column (200°C) and an electron capturedetector (ECD - 300°C). Carrier gas flow (helium -1 ml/min.) withnitrogen was at approximately 60 ml/min. The approximate retentiontimes for putrescine, cadaverine and hexanediamine were 3.6, 4.4and 5.1 min., respectively.Quantitative analysis of putrescine and cadaverine levelspresent were calculated (Shimadzu C-R4A Chromatopac) from the areabeneath the corresponding signal peaks relative to the internalstandard peak and they were based on a calibration curve generatedusing the dihydrochloride salts of putrescine and cadaverine(Sigma).Putrescine and cadaverine stock solutions were prepared bymixing 190 mg PUT-dihydrochloride, 170 mg CAD-dihydrochloride saltwith 0.1 N HC1 and the volume in each case was made up to 100 ml.Stock solutions were diluted with 0.1 N HC1 to yield working85solutions in the range of 1.0 to 160.0 ug/ml. The foregoing methodwas based upon the procedures of Staruszkiewiez and Bond (1981) asmodified by Farn and Simms (1987) and DFO (1989).Herring meal was spiked with working dihydrochloride saltsolutions of putrescine and cadaverine.Histamine determinationHistamine levels were determined by the fluorometric AOAC(1989) procedure based on the method of Staruszkiewiez et al.(1977). Briefly, 10.0 g of herring were homogenized (Polytron -2min) with methanol (50 ml) and the homogenate was made up to 100 mlwith methanol washings. The mixture was filtered and the filtratewas stored at 4°C.The sample clean up involved an anion exchange column (Bio-RadAG 1-X8) where 1.0 ml filtrate was loaded onto the column andwashed with distilled water until 35 ml of eluate were collected.The eluate containing histamine was made up to 50 ml with distilledwater and this was then stored in the refrigerator (1 week max.).The eluate (5.0 ml) was mixed on a vortex mixer upon each additionof 10 ml 0.1 N HC1, 3.0 ml 1.0 N NaOH, 1.0 ml o-pthalicdicarboxaldehyde (OPT) solution (100 mg OPT in 100 mldistilled in glass NeOH) and after 4 min. it was mixed with 3.0 ml3.57 N H 3PO4 . Thereafter the fluorescence intensity (I) (TurnerFluorometer- excitation wavelength of 350 nm and emissionwavelength of 444 nm) was recorded and compared with a standardcurve obtained with histamine dihydrochloride working solutions(0.5, 1.0 and 1.5 ug/5 ml).86Histamine was also determined using flow injection analysis(FIA) high-pressure liquid chromatography (HPLC) (reverse phase)analysis (Walters, 1984; Gouygou et al., 1987; Hungerford et al.,1990). Briefly, the histamine like substances were extracted fromthe sample (10.0 g) with methanol (50 ml mixed in Polytron and madeup to 100 ml with methanol washings).The filtered sample (Millex HV filters) extract was theninjected (25 ul) into the HPLC and derivatized on-line withorthophthalaldehyde (OPA) to form a flurophore. The HPLC systemconsisted of a 600E Multisolvent delivery system and controller,700 Satellite WISP autosampler, temperature control module #38000,post column reaction system and column heater chamber (Waters). AWaters 470 scanning fluorescence detector was used which monitoredexcitation and emission at wavelengths of 350 and 450 nm,respectively.A Novapack C18 column (15 cm long, 3.9 mm diameter) packedwith a particle size of 4 um, was used for reverse phasechromatography.Total volatile nitrogen and trimethylamine determinationTotal volatile nitrogen was measured by direct distillationwith MgO as described by Woyewoda et al. (1986) and bymicrodiffusion with K 2CO3 as described by Malle and Tao (1987).Trimethylamine was also determined by microdiffusion with K 2CO3 andformaldehyde as described by Malle and Tao (1987).874.3 RESULTS AND DISCUSSION4.3.1 Influence of spoilage on indicator levelsVolatile amines (TVN, TMA)The storage and processing conditions for the Pacific herringprotein sources are presented in table 4.The concentrations of TMA and TVN formed in the fish duringchill storage (2-5°C) before and after processing are presented inTable 5 and Fig. 5 (as is basis).The pattern of formation of TMA and TVN in bulk herring atchill storage was very comparable to that described in Experiment1. Little TVN and TMA were present initially. After 8 days, whichis the usual limit for acceptability of herring for humanconsumption (Fernandez-Salguero and Mackie, 1987), both the levelsof TVN(TVN-MgO) and TMA had increased substantially (Table 5). TheTVN values measured by direct distillation with MgO weresignificantly higher than those obtained by microdiffusion withK2CO3 and the values from both procedures were highly correlatedwith the TMA levels (r > 0.978).After 8 days of chill storage the majority of herring had"belly burst", since their gut walls were no longer intact(autolysis). Further, blood water had accumulated at the bottom ofthe storage container by day 12 (extreme autolysis).The levels of volatile amines (TVN, TMA) in the thawed wholeherring stored for up to 12 days at 2-5°C were similar to thosereported for non-frozen whole herring held on ice and were lowerthan those reported for whole herring stored at 5°C (Smith et al.,0g4- $.41 oa) $•r-i 0 .4.0440 trlg E-Ira $4 0$4W4 $.4(2)sa. -^1:4•HU) •a) Ea ti).r14-1 otil03 CIo S-1° wro Z(15 t7.10S-1gg cOrts w rcsgg U)u)>,.0 >7•UU) a) ...I$.4 '"4.1saeo (13c a)U)z.40 E •la 	 trl>4 a)to•o o--1w 004 f 1 U)4.) 0.) 4-)0 It0 C °U) W 04 01-1 ■-.1(1)4-1 (13Q) itare 	 t—t0 w• 4tn 	 4-)W b $.4r-i0 4-)4-) 4-4 T.1(C/O0sT04g88In N 	 v-I 	 0NM 105 CO r-I• • 	 • 	 • 	 •	 •	V LC)	 M %.0 	 CI 	 01	1-1 	 1-1 	 r l Cs1	MM 	 00 01 	 t-1	e--I	N 	 In	If) In 	 • • 	 • •• • 	 10 l0 	 01 01 	 1/40	■13	V' 'V 	 01 r-1r-I	CO N 	 10 r-I 	 1/40 CO 	 S	N 	 tn co co• •N ki3MMCs. N• •N VMMMM• •Cs) N•crCON N 01 01 1/40 0 N• • • • • •00 00 • r-I 0V CO CO M ■13 •4'N cNIIn CV ri N N 01 01CO CO Cl CO I-1 Ln 01• • • • • • •M 0 V N 0CO CO N CO 0 Mri CO1/40 CO CO CO In 1/40V V cr '1' CO 0 '1'• • • • • •CI CI r-1 1-1 N Co)CO CO N Nr-1 	 r1 CI 'VCO1-1 LC/ NN N Ntf1 cI• • • • • •V V V In k00 CON$4• •o4)N ADM r-4 0r-I 0 S M N Ol N• • 	 • 	 • 	 • 	 •riri 	 N10 	 CO 1-1CO 01CIM NCO 010NN 	 I' 	 IS M• • 	 • 	 • 	 • 	 •. V 	 CI V 1'	C*1 	 NIn If)CO 1/40 	 N 	 CON ri 	 CO In 	 O g:r 	 C.)• • 	 • 	 • 	 • 	 • 	 •11 rI 	 O T-1 	 M 0NN .4. CO MD N NCO CO 	 CO 0 	 CO 	 CO• • 	 • 	 • 	 • 	 • 	 •MCI 	 NCO 	 OtO 	 Cflto rnr-I N 	 U) l0II 	 II 	 II 	 NZ4.1 	 EZ 	 Cr) U)Eo 	 E-1 El 	 0Z 	 1.4 	 1.4$-4C.) 04-1ra4'CS>4 0.0 4'0 CU(U 04.)g 004.00 oto wvII1LS0g wrim4-)4-) a)00 ED4-I044-1C.) En1,4 Wa) EaU)a.)4-) w0>1 a,'CS CD4_3•r.1Wa.) cu$.4a)3 3 •0 ^0r—i 	 04W 04a) 4-1 °,--1 	 •LO;-.1c A• 4-10 	 U)(zI 0 a)C.) 0U) 	 r-10 034-1 >O89cc 	 coOOcc COO 0O 010 ■■■••■ ,• crztt 0CO 	 4-., .1 	.—.. 	 r.... ID 'g -0 . trCO 	 -,-1 4-3 aj C1 — 	 --1 4 .1-'1I (i)	 rrt a)1.0 	 4 "c"-ij P4) ,± 	rC) 0 an., =g (14 '''la (I-1 (1) • ■(/) 	 -r-I 4-) 7:5/*" 	 a.) '0 	 (I)ammlZ 0 tn,.. -f-1•f-I $.4 ).-. $.4a) ..-1 to 1.qt 	 > 0 co a)1 	 Ts >1 c.) a)ra 	 w NO XI 0 02 	 0 Pg),,I— 	 . g 	 4-4I 	 za)09-1 3ci "-,-1 03 ,-+ ,-,01 	 c.) ,-.1 	r,''''U) >4 'CI ^I 	 a) 4 C .I 	 )4 -P (3  -M 	4.) 0 	 I—IO g 4 L15i— 	 a-,-14.) 4 a)Ul 	 -IJCV 	 a) '1:1 - (az c E.,I	•f-1 ni 1-1° Z 	 (0LL 	 tt ..._ 1:1›-, g:, zF"' 	 r-I 	 (UI 	 0 . Tia c E.( wa) = r"14-1 	 --41.-- 	 0 F . 01 	 , r. aI 	 u) -..„ u) (f)r-I +:_, .,-4LL. 	 CD .-',.., tn 'CI1""" 	 > '"' (a C...I 	 CD w ,C1 (1::r--4 r-I-,-1 '0 0g 4-) a)al 	 4.4 -,-4a) 1 1:1 	Ti• 0 U) (1)> tal•IA r-ItU U) ..0• 4-1 I—I U)171 0 0 a)-ri 4-3 0) 3-4I , ...... 0 4.4901980; Fernandez-Salguero and Mackie, 1987). Once again thereappeared to be an increase in the lag period for the formation ofTVN and TMA in previously frozen herring stored at chilltemperatures. These findings are in agreement with those ofprevious studies (Luijpen, 1958; Shewan, 1961).A probable explanation for the foregoing findings is thatafter freezing and thawing the surviving bacteria are fewer innumber, metabolically injured and are more fastidious than theoriginal cells. Once the bacteria recover, however, spoilage occursat a rate similar to that occurring with non-frozen bacteria(Shewan, 1961; Hobbs and Hodgkiss, 1982). Borgstrom (1955)reported that bacteria in the vegetative state are sensitive to lowtemperatures and the colder the freezing temperature and longer theduration of storage, the greater the number of bacteria that arekilled. Christophersen (1968) agreed with these findings but hestated that while the micro-organisms may lose their ability tomultiply, they should not be considered dead since recovery of themetabolically injured cells is possible with peptides and othernutrients produced from enzymatic protein digests. Speck and Ray(1977) also warn that bacteria may only be injured but not killedas the result of frozen storage and spoilage micro-organisms can beefficacious if they are permitted to repair and grow duringthawing. Thus, the shelf life of frozen-and-thawed cod compared tothat for fresh cod was found to be extended if frozen storageexceeded one month. Once again, this was attributed to theincreased lag phase for bacterial production in the prefrozen cod91(Licciardello and D'Entremont, 1987).The trends for levels of non-volatile amines in the thawedherring paralleled those for the levels of the volatile aminesduring chill storage with the exception of histamine which remainedlow (16.3 ppm) even after 12.5 days (Table 5). The putrescine andcadaverine levels increased significantly after day 8 withsubstantial formation at 12.0 days of 50.9 and 524.8 ppm,respectively. Once significant microbial growth had occurred thelevels of putrescine and cadaverine increased exponentially as byday 12.5 the levels were 96.2 and 553.3 ppm, respectively. Thesefinding support the work of Ababouch et al. (1991), who reportedlarge increases in putrescine and cadaverine for sardines after 8days of storage on ice. They attributed the rapid increase in thelevels of amines to proteolysis which increased the availablelevels of free lysine and ornithine which were then decarboxylatedto cadaverine and putrescine.Putrescine and cadaverine levels formed during chill storageof whole herring compared closely to those previously reportedduring iced storage of non-frozen herring (Ritchie and Mackie,1980) but they are lower than the levels found for whole herringstored at 5°C (Fernandez-Salguero and Mackie, 1987) this is despitethe fact that the initial levels (putrescine, cadaverine andhistamine) were the same as those reported for deep frozen herring(Lebiedzinska et al., 1991). This observation suggests that pre-freezing of herring before chill storage may also increase the lagperiod prior to significant microbial growth and associated92putrescine and cadaverine formation.The cadaverine levels found in the present study are higherthan those noted by Klausen and Lund (1986). These authors reportedcadaverine concentrations of 20 ppm and 40 ppm in herring storedfor 15 days at 2°C and 10 days at 10°C. This difference may beattributed to the use of fresh, eviscerated herring in the study byKlausen and Lund (1986). Evisceration of herring before storagewould infuence the formation of biogenic amines since a large partof biogenic amine formation occurs in the belly walls and withoutthe viscera there would be a lower production of cadaverine.Putrescine levels in spoiled fish are generally much lowerthan those of cadaverine, probably because of the limitedquantities of free ornithine present in the fish muscle (Taylor andSumner, 1987). Thus the initial low putrescine levels in thepresent study may reflect the low initial free ornithine levels,but once autolysis begins (after day 8), the levels of putrescineincreased rapidly which may reflect the elevated levels of peptidesand free amino acids from the autolytic process that have beenreported by De Silva and Hughes (1962) and Aksnes (1988; 1989).Hence, gutting fish as soon as possible will prolong shelf life(Fernandez-Salguero and Mackie, 1987; Connell, 1990).The histamine levels remained low in the herring duringstorage at 2-5°C for the duration of the experiment (12.5 days).This contrasts with the findings of several investigators who havereported high histamine contents (> 100 ppm) for herring stored atchill temperatures (Ritchie and Mackie, 1980; Klausen and Lund,931986; Fernandez-Salguero and Mackie, 1987). A possible explanationfor the descrepancy in results is that the herring were storedfrozen (-20°C) for up to three months before thawing and chillstorage. Investigators have reported that pre-freezing of fishinhibits the rate of subsequent histamine formation (Baranowski andPan, 1985; Baranowski et al., 1990). Baranowski et al. (1990)reported that the longer the freezing time the greater thedepression of histamine levels during subsequent storage. Moreover,they found that inhibition of histamine formation was greatest inmahimahi that had been stored for 40 weeks at -20°C andsubsequently incubated at 32°C (12 ppm histamine) relative to non-frozen controls which had been incubated at the same temperature(2920 ppm histamine). Van Spreekens (1987) reported similarfindings for herring, since bacterial populations were greatlyreduced after frozen storage and the lag period for histamineformation in thawed herring stored at 4°C and 7°C lasted for 9 and6 days, respectively.Smith et al. (1980) reported low histamine levels (< 7.0 ppm)for herring caught in February and July and stored in ice,refrigerated sea water and air (ambient temperature). By contrast,herring caught in November and stored under the same conditionsproduced high levels of histamine (156 ppm). It is possible thatlow histamine production was due to reduced histidinedecarboxylating bacterial numbers and initial levels of histidinein the fish before storage. In support of this, both bacterialpopulation numbers and species and/or histidine levels are known to94fluctuate seasonaly in herring (Hughes, 1959; De Silva and Hughes,1962).Histamine is usually formed later than putrescine orcadaverine and eventually its concentration reaches a higher levelthan those for other amines in herring (Ritchie and Mackie, 1980;Klausen and Lund, 1986; Fernandez-Salguero and Mackie, 1987).Histamine levels did increase slightly by day 12.5 but it is notknown if further storage at 2-5°C would have resulted in theformation of even greater concentrations of histamine in theherring.The low levels of histamine (< 5.0 ppm) obtained by theA.O.A.0 method (1989) in both the herring and herring meal wereconfirmed by high performance liquid chromatography (Table 5).4.3.2 Influence of thermal processing on indicator levelsTotal Volatile Nitrogen and TrimethylamineThe levels of TVN were lower in all the presscake meals thanin the raw material, with the greatest losses occurring in the mostspoiled meals. The TVN levels were lower in the high temperaturemedium and spoiled meals (HTMH-4, HTMH-6) when compared to theirlow temperature counterparts (Tables 5, 6).The TVN levels were significantly higher in the presscakemeals when the direct distillation with MgO method was used ascompared to the microdiffusion with K 2CO3 method. The trend for TVNincrease in the raw material and processed meals was very similarwhen both methods were used (Table 5). These findings are inagreement with those in experiment 1. The higher TVN values for the95Table 	 6. 	 Retention(%) 	 of 	 spoilage indicators 	 (level	 afterprocessing/level before processing) 	 x 100) 	 in the herring presscake meals 	 after processing 	 (Referinformation).to Table 	 4 for additionalSample Put' Cad 	 Hist TVN-MgO TVN- TMA-K2CO3 K2CO3LTFH-1 46.6 27.7 	 - 98.1 79.9 218.9HTFH-2 30.4 24.6 	 - 99.2 90.2 250.7LTMH-3 10.8 8.3	- 65.8 46.7 29.2HTMH-4 16.3 20.1 	 - 66.5 42.9 35.9LTSH-5 17.8 16.9 	 50.0 30.9 22.4 15.9HTSH-6 10.9 16.6 	 40.1 23.7 18.3 19.3FDH-7 36.1 52.0 	 - 97.1 97.4 147.7'Abbreviations: Put, putrescine; Cad, cadaverine; Hist, histamine;TVN-MgO, total volatile nitrogen as determined by directdistillation with MgO; TVN-K 2CO3 , total volatile nitrogen asdetermined by microdiffusion with K 2CO3 ; TMA-K2CO3 , trimethylamineas determined by microdiffusion with K2CO396direct distillation with MgO may be the result of protein breakdownas discussed by Botta et al. (1984). Both methods for TVNdetermination ranked the raw material and the presscake meals inthe same order (fresh to spoiled). Therefore, either method may beused for assessing spoilage in the raw material but sincemethodology is a factor, care should be taken when comparing theTVN results between studies.Gallardo et al. (1990) found that the levels of TVN, TMA, DMAand ammonia increased while those for trimethylamine oxide (TMAO)decreased during precooking and canning of tuna. Extended heatingtimes and elevated processing temperatures led to increased levelsof ammonia, TMA, DMA and TVN. Cooking herring at 120°C resulted inthe breakdown of TMAO and formation of ammonia, TMA, DMA and mono-methylamine (Hughes, 1959). Moreover, the amount of amines producedwas directly proportional to the cooking time.Thus, TVN would be expected to increase in the herring duringprocessing since TVN is composed mainly of TMA, ammonia and to alesser extent DMA and monomethylamine. The thermal decompositionof TMAO to TMA and DMA takes place even at 55-60°C but formation isgreatly accelerated at higher temperatures and by the addition ofsugars (Tokunaga, 1975). Therefore, the decrease in TVN levelsnoted in the presscake meals in this study seems to contradict theprevious findings.Toyama et a/. (1981) reported that the levels of TVN werehigher in whole fish meals (solubles added back to meal) whencompared to press cake meals. Upon further investigation, Toyama et97al. (1982) found that the levels of TVN increased during thermalprocessing of the raw material but the TVN was concentrated in thesolubles and stickwater during the cook and press process. Theyalso found that when TVN levels were low (fresh material) most ofthe TVN was retained in the presscake meal. However, when TVN washigh in the raw material (spoiled) most of the TVN was present inthe stickwater after processing. Increasing the drying temperaturefrom 80 to 120°C resulted in a decrease in the TVN levels in boththe presscake and whole fish meal (Toyama et al. 1982). Sandfeld(1983) also reported an increase in TVN levels in presscake mealsand whole meals when the raw material was spoiled but TVN levelswere decreased in overheated presscake meals produced from spoiledraw material. He rejected TVN as a good indicator of meal qualitysince most of the TVN remained in the solubles and press liquor(stickwater). Thus, TVN is used as an indicator of spoilage in theraw material. With respect to this, good quality fish meals areproduced from fish with TVN levels of < 90 mg N/100 g fish (Pike etal., 1990).The increase of TMA levels in the fresh presscake meals waslikely due to the thermal decomposition of TMAO which is initiallyhigh in the fresh material (Hughes, 1958; 1959). During spoilagethe TMAO is reduced to mostly TMA by bacterial triamine oxidasewhich results in lower TMAO levels in the raw material. Hence, ifthe TMAO has already been converted to TMA, the now available TMAmay also be lost in the solids and press liquor during the cook andpress operation. However, increased processing temperature (100°C)98during meal production in the present study resulted in elevatedlevels of TMA and a larger retention of TMA in the presscake mealsas compared to low temperature processing (75°C) (Table 5,6).The foregoing results may be explained by the findings ofprevious investigators who have reported that TMA and DMA levelsincrease during thermal processing. Sigurdsson (1947), forinstance, reported an increase in TMA levels during the retortingof herring. Hughes (1958; 1959) also found that TMA levelsincreased from 5.1 to 18.3 mg N/100 g fish after herring wereheated at 120°C for 1 hour and the levels increased further to amaximum of 63.4 mg N/100 g fish after 4 hours of heating. Similarlythermal processing of tuna led to elevation of TMA levels in theraw material from 280 to 340 mg N/kg during precooking (102°C) andto 450 mg N/kg fish during retorting (110°C) (Gallardo et al.,1990). Tokunga (1975) also reported increased TMA levels in variousmarine species of fish during thermal processing at 100°C andfurther TMA increases at 120°C.By contrast, Suzuki et al. (1987) reported a decrease in TMAlevels in the dark meat of tuna as processing temperaturesincreased from 100 to 120°C and a similar but lower loss in thewhite meat.The increase of TMA levels noted during thermal processing ofthe herring in this study may have resulted from water lossesduring processing which would have increased TMA levels on a wetmatter basis. Therefore comparisons of indicator levels in raw andprocessed product should be expressed on a dry matter basis in case99of any moisture losses during processing.The levels of TMA increased in the herring during freeze-drying (Tables 5,6). Freeze-drying has been shown to convert TMAOto TMA and DMA in fish products (Kida and Tamoto, 1976 as cited byHebard et al., 1982). Spinelli and Koury (1979) reported that drum-drying and freeze-drying hake greatly increased DMA and TMA levels.Slabyj and True (1978) suggested that the TMA test should beperformed on raw sardines to estimate the quality of the cannedproducts. Perhaps the TMA test should also be performed on the rawfish to estimate the quality of the fish meal.Putrescine, Cadaverine and HistamineThermal processing decreased the levels of non-volatile amines(putrescine, cadaverine, histamine) in the presscake meals (Table5). Presscakes with high initial levels of amines in the rawproduct still retained high levels of amines relative to the otherpresscake meals even though the press liquor and some solids werenot added back to the presscake meal (Table 6).The retention of the volatile amines was higher in thepresscake when initial amine concentrations were lower in the rawmaterial. The retention of non-volatile amines was low in thepresscake meals. The ranking order of the meals based on the levelsof non-volatile amines (putrescine, cadaverine) remained the sameas for the raw fish. In general the high temperature meals (100°C)had lower levels of non-volatile amines when compared to the lowtemperature meals (75°C) except for HTMH-4. It is possible thatputrescine and cadaverine levels were underestimated in the raw100herring for HTMH-4.The non-volatile amines are very water soluble and aconsiderable amount of histamine has been shown to be pressed outby the "cook and press" operation during fish meal production(Toyoma et al., 1981). Toyoma et al. (1982) discovered that eventhough histamine levels increased in the raw material duringstorage, they decreased during cooking and drying (80-100°C), andmost of the histamine was retained in the stickwater and solubles.Watanabe et al. (1991) confirmed these findings by showing that thelevel of histamine in the "water extract" produced during fish mealproduction paralleled the levels of histamine formed in the rawmaterial during storage.Also, elevated cooking and drying temperatures during mealproduction result in enhanced amounts of press liquor and solublesand the increase is greater if spoiled raw material is employed(Ward et al. 1977). Thus, more press liquor (containing putrescine,cadaverine and histamine) would be produced at 100°C from thespoiled herring (HTSH-6) than from the fresh herring processed at75°C. Since most of the solubles were not added back to thepresscake meal in this study, the results provide information onamine levels that would be near their lowest possibleconcentrations in herring meal in relation to time of storage ofraw material.Histamine levels in fish have been shown to decrease withincreased thermal processing (80-120°C) due to the non-enzymaticbrowning reaction (Maillard reaction) as well as binding of101histamine with protein (Tanaka et al., 1986). Similar decreases inputrescine and cadaverine levels have been observed in tunasubjected to pre-cooking (71-82°C) and further reductions inputrescine and cadaverine levels have been found after retorting at117°C (Farn and Sims, 1987). The reduction in putrescine andcadaverine during thermal processing of fish is expected, sinceduring heating cyclization of putrescine gives rise to pyrrolidine(pyr) while cadaverine gives rise to piperidine (pip). However, theexact amount of putrescine and cadaverine converted to pyr and pipremains unclear (Shimakura et a/., 1991).The levels of pyr have been shown to increase in raw sardinesduring storage at 10°C and during fish meal production whenprocessing temperatures range from 80 to 140°C. Pyr levels in thefish meal were noted to decrease with heating at 160°C and when thefish meal was stored at 10 to 30°C for up to 30 days. Although, pyrlevels remained constant in fish meals stored at - 10°C for up to30 days (Tozawa and Kawabata, 1987).Pike et al. (1990) also reported decreased cadaverine,putrescine and histamine levels in fish meals dried at 140°Crelative to meals dried at 60°C.Maximum levels of putrescine (10.5), cadaverine (91.9) andhistamine (6.52 ppm) were obtained in the high temperature spoiledherring presscake meal (HTSH-6). The levels of these non-volatileamines were much lower than those reported by Pike et a/. (1990)(putrescine-790, cadaverine-4500 and histamine-4830 ppm) for wholefish meals processed from spoiled raw material. The holding102temperature, duration of storage time and nature of raw fish werenot given for the latter experiment but the raw material must havebeen very spoiled for such high histamine and cadaverine levels tobe recorded. The maximum levels for the amines even in the mostspoiled meal (HTSH-6) for the present study were still lower thanthose reported by Smith (1990) for Danish fish meals. In thisregard, the maximum levels (ppm) of putrescine, cadaverine andhistamine in special and standard Danish fish meals wererespectively (25 to 61), (64 to 117) and (10 to 20). However, thelevels of putrescine, cadaverine and histamine in the spoiled lowand high temperature presscake meal (LTSH-5, HTSH-6) were higherthan those values found for the commercial Pacific herring mealused in the present study.The most useful index which was found for assessing spoilageof the raw material and the fish meal in the present study wascadaverine. Since cadaverine levels increased the most in both themedium and advanced spoiled raw herring and herring meal. Indeed,cadaverine levels appear to be the most useful index for evaluatingspoilage in fish regardless of the species (Yamanaka et a/., 1986;1987). Other investigators (Mietz and Karmas, 1977; Rosier andPeteghem, 1988) have used a chemical quality index based upon theratio of histamine, putrescine and cadaverine levels to those forspermidine and spermine in fish to monitor decomposition. Morerecently, Lebiedzinska et a/. (1991) reported that assessment ofthe quality of herring processed in various ways (frozen, saltedand smoked) could be accomplished by examining the levels of103biogenic amines (histamine, cadaverine, putrescine, tyramine,spermidine and spermine) using multiple principal componentanalysis (multivariate statistical method). This approach confirmedthat the method of fish processing is crucial for both the contentsand composition of biogenic amines in the finished product.The GLC procedure employed for the determination of putrescineand cadaverine was applied to extracts of fresh herring presscakemeal spiked with the dihydrochloride salts of putrescine andcadaverine at two levels (Table 7). The average recovery forcadaverine was 99.7% (range 96.6-101.5%) for single determinations.For putrescine the average recovery was 94.5% (range 89.3-99.5%)for single determinations. These recoveries are in close agreementwith the recoveries of putrescine and cadaverine from spiked tunareported by Staraszkiewicz and Bond (1981) and recoveries reportedby DFO (1989) of 95% or better for cadaverine and 80-90% forputrescine from spiked tuna. A representative chromatogram is shownin Fig. 6.4.4 CONCLUSIONSThe levels of putrescine and cadaverine increasedsubstantially in Pacific herring after 8 days storage at 2-5°C.Cadaverine was produced at high enough levels to possiblypotentiate histamine toxicity. Although, histamine levels remainedlow (< 15 ppm) in herring even after 12.5 days of storage. Aprobable explanation for this is that the long term frozen storageof the herring used in the present investigation before chillstorage resulted in an increased lag period for the formation of104• roa) a)q• z1O 7:$Utx '44"a-)za)10 	 r0CI) rt,C.) 	 $4 'CI(I) (150 0(U	0(1) $.41:4 4-10'0a)4.) co(c)▪ 0(CE-1WO "1021rCS04 (171	 1 04rr:J 0r0rCITS44 -,▪ -40 z 	 >W 0—I ;4W0	tin 	 A11-1• tp• •H 	 a)c:24r-I 	 0.1A RI 	 Cl)WEl 0Cel 1/40 	 h 	 011.11 0 CV tNO ocrl cm co a%h- C101 01.CV CN N NIn In to• • 	 • 	 •	I -I 	 1-1H N C1 .4.toH Inri 	 rI0 0r-I 	 1-1H r-I r i H0 0 0 0• • 	 • 	 •H H H HH N M .4.•qt'0	 1-♦•-•coCV	 Co00co	 cr%CO 	 r-ICI 	 1/4001 	 1/4001 01InM CO01 CO01 01•:1"01 01Hre)105Cr'T-CJ‹LC\CoA (=>L 	 L 	 L 	 L 	 L Retention Time (min)Fig. 6. Gas chromatogram for standard calibration mixture ofbiogenic amines. Peak identification and retention times: PUT,putrescine 3.62 min; CAD, cadaverine 4.43 min; HEX, hexanediamine5.11 min. (internal standard).106histamine. It is unknown whether or not pre-freezing increased thelag period for the formation of putrescine and cadaverine as well.The levels of volatile amines (trimethylamine and totalvolatile nitrogen) also increased during chill storage with themajority of TMA and TVN production occurring after day 12.Thermal processing increased the levels of TMA in the freshand freeze-dried herring meals but decreased the levels of TVN andTMA in the remaining presscake meals in relation to initial levelsin the raw material. Some of the TMA and TVN was lost in the pressliquor which was not added back to the meal. The levels ofputrescine, cadaverine and histamine also decreased duringprocessing and probalby some of the non-volatile amines were lostin the press liquor and also they may have been converted topyrrolidine and piperidine. The higher the processing temperaturethe greater the conversion of TMAO to TMA, especially in the mealsproduced from the fresh raw materials. An increase in processingtemperature decreased the levels of putrescine, cadaverine andhistamine in the fish meals.The simultaneous determination of putrescine, cadaverine, andhistamine in herring is possible with gas-liquid chromatography orhigh performance liquid chromatography. The assessment of therelationship between the levels of biogenic amines in herringduring storage and processing may prove to be useful for predictingthe quality of fish meal. The determination of TVN and TMA levelsin the raw product as indices of advanced spoilage is recommendedbefore fish meal production.107Chapter 55.0 EXPERIMENT 3. Influence of raw material freshness and ofprocessing conditions on the nutritive value of Pacificherring (Clupea harengus pallasi) for rainbow trout(Oncorhyncus mykiss) in fresh water and chinook salmon(Oncorhyncus tshawatscha) in salt water.5.1 INTRODUCTIONPrevious research has been shown that the nature of the rawmaterial and the processing temperatures employed during fish mealproduction influence the quality of marine fish protein andconsequently the dietary nutrient and energy digestibility andgrowth performance of chickens and salmonids (Tarr, 1982; Sandfeld,1983; Opstvedt et al., 1984; McCallum and Higgs, 1989; Pike et al.,1990). Spoilage of the raw material, before fish meal productionresults in elevated levels of histamine in fish meals (Toyama etal. 1981; 1985). Diets containing fish meal with high levels ofhistamine have been found to cause stomach erosion and ulcerationwhen fed to rainbow trout and chickens (Watanabe et al., 1989).Overheated (>100°C) fish meals have been found to havedepressed nutritive value relative to low temperature dried (60-75°C) meals on the basis of the poor growth performance ofsalmonids fed diets containing such meals (Sandfeld, 1983; McCallumand Higgs, 1989). Moreover, Andorsdottir et al. (1989 as cited inPike et a/., 1990) reported that overheated fish meal had a lowerprotein and amino acid digestibility in Atlantic salmon than lowtemperature dried fish meal. Pike et al. (1990) also found thatmeals dried at 60°C from stale fish had 2 to 3% lower proteindigestibility in mink than meals made from fresh fish. When the108meals were dried at 140°C digestibility differences between freshand stale raw materials were not found. The meals made from stalefish had higher pepsin solubility values (in vitro digestibility).However, it is not known how processing temperature and spoilage ofthe raw material will affect protein and energy digestibility insalmonids and few investigators have examined the interactiveeffects of these variables on fish meal quality.The aims of the present study were therefore to determine theeffect of three levels of spoilage and two processing temperatureson the potential nutritive value of Pacific herring meal forrainbow trout in fresh water and chinook salmon in salt water.Potential nutritive value was assessed by both in vitro tests(pepsin and pH-stat test) and in vivo (modified Guelph system)digestibility tests and the values were considered in relation tothe levels of amine compounds in the meals. It was hoped that theinformation obtained would prove to be useful in predicting thequality of fish meal.5.2 MATERIALS AND METHODS5.2.1 Test protein sourcesThe protein sources were prepared from ripe, male Pacificherring (Clupea harengus pallasi) as described in Experiment 2(Fig. 7)(section 4.2.1, Table 4). A two by three factorialexperiment was designed with two drying temperatures and threelevels of raw material freshness to investigate the influence ofraw material freshness and processing temperature on the nutritivevalue of the press cake meals. A freeze-dried control proteinS. CloncyFishMealTemperatureProbeSteam-jacketedCooker•ExhaustTemperatunProbeFish ProductSteam-jacketed RotaryDryer (Indirect)109Fig.1 Pilot Fish Meal Manufacturing Machine (Chemical Research Organization, Denmark)110source was also produced from fresh, frozen fish.5.2.2 Diet composition and preparationThe proximate composition and gross energy contents (Table 8)of the test meals were determined as follows; moisture (20 hrs at85°C), ash (2 hrs at 600°C), crude lipid (Bligh and Dyer, 1959),total nitrogen (micro-Kjeldahl, Technicon Auto Analyser) and grossenergy (adiabatic bomb calorimeter-Parr Model 1241). The amino acidcomposition of each of the test meals was determined by Amino AcidAnalyser (Beckman-Spinco Analyzer, Model 120B with DionexAnalyzers, Models D-500 and D-502) at Triple A Laboratories,Seattle, U.S.A. (Table 9a,b). The mineral compositions (Table 10)were determined by plasma emission spectroscopy (ICAP) at QuantaTrace Laboratories Inc., Burnaby, B.C. Spoilage and lipidperoxidation indices measured for each of the test protein sourcesincluded total volatile nitrogen (TVN), trimethylamine (TMA),putrescine, cadaverine and histamine as determined in Experiment 2.Peroxide value (PV) was determined as described by Woyewoda et al.(1986). The spoilage indicator levels are presented in Tables 5 and11.The composition of the basal diet is provided in Table 12.Each of the mineral supplements were ground to a fine powder in acoffee grinder (Braun) before they were weighed individually andmixed with ground wheat middlings as the carrier in a twin shelldry blender (V-mixer) for 45 minutes. The same process was followedfor the vitamin supplements but pregrinding was not necessary. Allfeedstuffs were ground in a Fitzmill (Model JT) equipped with a111Table 8. Proximate compositions and energy contents of test herringmeals (values in parentheses are on a dry matter basis).Proximate constituent (%)CodeWater Ash CrudeproteinCrudelipidG.E.(MJ/kg)LTFH-1 14.5 9.5 62.7 13.6 19.9(11.1) (73.3) (15.9) (23.3)HTFH-2 16.5 9.6 61.5 12.9 19.4(11.5) (73.7) (15.5) (23.2)LTMH-3 12.1 10.3 63.8 14.2 20.7(11.7) (72.7) (16.2) (23.5)HTMH-4 11.5 10.1 66.1 13.0 20.8(11.4) (74.7) (14.7) (23.5)LTSH-5 12.8 12.6 63.7 11.0 19.5(14.4) (73.0) (12.7) (22.4)HTSH-6 12.0 12.8 63.0 12.3 20.1(14.6) (71.6) (14.0) (22.8)FDH-7 5.1 11.0 65.5 18.9 23.0(11.6) (69.0) (19.9) (24.2)REF-8 6.8 13.4 69.3 11.0 20.5(14.4) (74.3) (11.8) (22.0)IngredientNameLow temperaturedried freshherring meal(HM)High temperaturedried fresh HMLow temperaturedried mediumspoiled HMHigh temperaturedried mediumspoiled HMLow temperaturedried spoiled HMHigh temperaturedried spoiled HMFreeze-driedfresh raw herringCommercialHMTable 9a. Amino acid composition of herring meal protein sources' (Refer to table 8information)for furtherAMINO ACIDS LTFH-1 HTFH-2 LTMH-3 HTMH-4 LTSN-5 HTSH-6 FDH-7Alanine 4.05 3.99 4.11 3.95 4.03 3.58 3.80Arginine 7.13 6.66 6.34 7.36 6.19 5.89 6.55Aspartic acid 6.16 6.28 6.14 6.38 6.87 5.98 5.54Cystine2 0.51 0.40 0.62 0.58 0.55 0.55 0.57Glutamic acid 8.88 8.68 8.65 8.82 9.76 8.25 8.21Glycine 4.65 4.28 4.56 4.20 4.23 3.64 4.84Histidine 1.40 1.43 1.39 1.40 1.40 1.31 1.15Isoleucine 3.03 3.28 3.05 2.84 3.41 2.73 2.64Leucine 5.23 5.88 5.15 5.24 6.18 5.44 4.95PLysine 5.26 5.47 5.56 5.33 5.37 5.13 4.56 HNMethionine 1.86 1.93 2.04 1.56 1.90 1.84 1.72Phenylalanine 2.98 2.83 2.75 2.86 3.00 2.92 2.42Proline 3.63 3.16 2.68 3.13 3.44 1.61 3.51Serine 3.19 3.26 3.13 2.97 3.58 3.00 3.00Threonine 3.04 3.07 4.13 2.95 4.76 2.82 3.37Tryptophan3 0.71 0.71 0.66 0.71 0.68 0.75 0.51Tyrosine 2.35 2.42 2.33 2.41 2.56 2.36 1.99Valine 2.87 3.87 3.83 3.60 3.42 3.24 3.46Total 66.90 67.54 67.12 66.33 71.33 61.02 62.80Expressed as g/100 g dry weight sample. As determined by Triple A Laboratories, Seattle, WA2 Calculated from cysteic/alanine ratio.3 48 hour alkaline hydrolysis @ 135°C.Table 9b. Amino acid composition of herring meal protein sources l (Refer to table 8information)for furtherAMINO ACIDS	 LTFH-1 HTFH-2 LTMH-3 HTMH-4 LTSH-5 HTSH-6 FDH-7Alanine 	 5.52 5.41 5.65 5.29 5.51 5.00 5.51Arginine 	 9.73 9.03 8.72 9.85 8.48 8.22 9.50Aspartic acid 	 8.41 8.51 8.45 8.55 9.41 8.35 8.03Cystine 2 	0.70 0.55 0.85 0.78 0.76 0.77 0.83Glutamic acid 	 12.11 11.78 11.89 11.81 13.37 11.52 11.9Glycine	 6.35 5.80 6.28 5.63 5.80 5.08 7.01Histidine 	 1.91 1.93 1.91 1.88 1.92 1.83 1.66Isoleucine 	 4.13 4.45 4.19 3.80 4.67 3.81 3.83Leucine 	 7.13 7.98 7.09 7.02 8.47 7.60 7.18I-.Lysine 	 7.18 7.43 7.65 7.14 7.35 7.16 6.61 HwMethionine 	 2.54 2.62 2.80 2.09 2.61 2.57 2.49Phenylalanine	 4.07 3.83 3.79 3.83 4.12 4.08 3.51Proline	 4.95 4.29 3.69 4.19 4.71 2.25 5.09Serine 	 4.36 4.42 4.30 3.98 4.90 4.19 4.35Threonine 	 4.15 4.16 5.68 3.95 6.52 3.94 4.89Tryptophan3 	0.97 0.96 0.90 0.95 0.93 1.04 0.74Tyrosine 	 3.21 3.28 3.21 3.22 3.50 3.30 2.89Valine 	 3.91 5.25 5.27 4.83 4.68 4.52 5.01Total 	 91.27 91.64 92.32 88.79 97.77 85.22 91.01Expressed as % protein of dry weight sample. As determined by Triple A Laboratories, Seattle,2 Calculated from cysteic/alanine ratio.3 48 hour alkaline hydrolysis @ 135°C.114Table 10. Mineral composition of herring test protein sources''ELEMENTANALYSIS ,LTFFP-I HTFH-2 LTMH-3 HTMH-4 LTSH-S 1.ITSM-6, FDH-7,Calcium 29350 29700 28350 30500 31400 31450 21500Chromium 7.22 6.82 8.66 5.12 6.99 4.84 4.98Cobalt < 0.2 < 0.2 < 0.2 < 0.2 < 0.2 < 0.2 < 0.2Copper 10.55 6.35 7.79 7.83 6.90 4.95 3.30Iron 984.0 768.0 967.0 797.0 866.0 686.0 105.0Magnesium 2035 2035 1790 1935 1480 1545 1755Manganese 15.0 13.5 14.5 14.0 12.0 11.0 7.35Phosphorus 25700 25750 23300 25100 22200 22400 21650Potassium 8490 7995 6295 6340 5450 5280 11550Selenium < 0.2 < 0.2 < 0.2 < 0.2 < 0.2 < 0.2 < 0.2Sodium 7670 7090 5510 5640 19750 20750 9950Zinc 95.1 96.4 91.9 97.8 89.1 87.5 71.5I Expressed as mg/kg dry meal as determined by Quanta Trace Laboratories Inc.Burnaby. B.C.O(aU)0•(I)C.)4_)O 1/O(U) )-4z o.4-1 4'1z.6)r"• -I04 al.4J 0U)W 4-)fUf--4 rt:▪ $tttr) 0• 4-1• co$4a)• 4E-14-)$.4 0O 4-)W• 4-1O W4.) p4C.)•,-1 	 •U)U)4-1 >I0U)a) (11a) 0. r00r4,C1 04• XE.+ (1)115O1 N 4:14 If) 0 4:1' 0 10V I-I M N 10 N r1dr 0•01•1-1•LC)•V' CO•d'CN N r1 CV rl r-1 N 1-1In 	 CV 	 10 	 1/40 	 1-1▪ M 	 N CO ri• •	 •tn M %.0 to elr-I 	 1-1 	 I-1 	 r-I 	 r-1 	 CNCO N V) 4-.1 tO CO N T-IN N N LC) CO CO lf)• • 	 • 	 • 	 • 	 • 	 •CV 1/40 N d' N N 01 01M el c.1 CI1•1 0^) 	 In CV IN V)IN V) IN ■D Cs) 	 r') 0• • 	 • 	 • 	 • 	 • 	 • 	 •N▪ 10 C•1r-I 0 N 01r-I 	 ri 	 CNI.CO 10 If) CV 101-4 CO r) 0• •1•1 	 rmi 	 0 	 1-1M d' 	 kor•• 	 COI44 C441 H E"4 	 l	 4144 c4(ari.4o 4r--1Oa)r-4• 1to	 Ts	 tio-1 	 Q),S4a)	 zg• (I)	 0	 a)•ro	 $4	 04a)> 0 	 a)04	 4-)	 a)4 	 rcs04 	 b• MI• (-) 	 0triO g $44-)o W 	 4.)P tr) 	 g4 alU r-1fa 	 rofi) 0 	 t--1lo 04 	 >1tn rn 	 . 	 04-) 	01(1) 	 W,r) 0 	 .,..,4-41-4a)• Ln 	 4-) 	 • r4-I-1 a) 	 TS43• ra4.) H	Pozrt5 04-1 	E-I	 U)a)g )-)$.4 4-4 	a)w w 0+• g ob rm.	 ••••■■al /4	 C.14•a) 6,4:1 	 e,1 f••••1.pO0 4-3 WrC)W 1-1• (1) 	tU	 o4-)	 4_) o• W 	 O -4 a)04 ti • 4-) u) 041,1.4 W 	 4 4_1 4-1 (4.4O Z 010 4-1 oU) 	 N '0 U)W 0 ,1• UIw 4-) 4-) (1) $-1 WC.)Q1 "--1 	Q)	 Q))4 4 	 14 	 ,4OO0)116Table 12. Composition of basal mix used for preparation ofreference and test diets (70% reference : 30% testingredient) employed in experiments III and IV.Ingredients Concentration(g/kg dry diet)Pacific herring meal 385.0Poultry-by-product meal 70.0Blood flour 40.0Corn gluten meal 65.0Dried whey 60.0Wheat middlings 186.6Vitamin supplement' 20.0Mineral supplement 2 30.0Herring oil 3 108.2Choline chloride (60%) 5.0Ascorbic Acid 2.0DL-methionine 3.2Permapell (lignin-sulphonate binder) 10.0Finnstim (betaine-appetite enhancer) 15.01 Vitamin mix provided the following amounts per kg of dry diet:inositol 400 mg; niacin 300 mg; pantothenate (as D-calciumpantothenate) 165 mg; riboflavin 60 mg; pyridoxine (as pyridoxine-HC1) 40 mg; thiamine (as thiamine mononitrate) 50 mg; menadione 18mg; folic acid 15 mg; biotin 1.5 mg; vitamin B u 0.09 mg; vitamin E300 IU; vitamin D3 2400 IU; vitamin A 5000 IU; BHT 22 mg with 18.79g of wheat middlings (carrier).2 The mineral mix supplement provided the following (mg/kg of drydiet): P (as KH 2PO4) 1881; Mg (as MgSO471120) 933; Cu (as CuSO 4 .5H20)6.33; Fe (as FeSO 4 .7H20) 38.2; Zn (as ZnSO47H20) 104.9; Mn (asMnSO4 .H20) 75.4; Na (as NaC1) 1145; K (as K 2SO4) 3025; I (as KI0 3 )5.0; F (as NaF) 4.5; Co (as CoC12 -6H20) 2.80; Se (as Na 2SeO3) 0.20;Al (as A1C1 3 . 6H20) 5.0 and 6.62 g of wheat middlings (carrier).3 Stabilized with .05 % Santoquin (antioxidant)117size 30 U.S. screen (0.595 mm) before they were mixed with thevitamin and mineral supplements. The dry components were mixedthoroughly and then a portion of the fish oil was added. Totalmixing time was 45 min. in a Marian mixer. After preparation of thebasal mix, 5 kg batches of the reference (contained 0.5% chromicoxide as indigestible marker) and test diets (70% reference: 30%test fish meal with 0.5% chromic oxide) were prepared in acommercial Hobart mixer (model M-802) (mixing time was 30 minutes).Special care was taken to ensure that chromic sesquioxide(Cr2O3) (Fisher Scientific) was mixed uniformly into the referenceand test diets. In this regard Cr2O3 was mixed with a portion ofeach of the test diets and reference diet at a level of 0.5% usinga Hobart mixer for 20 minutes followed by a V-mixer for 30 minutes.All meals were adjusted to 9.0% moisture before final mixing andthen pelleting.The individual diets were then pelleted in a California PelletMill using a 3.18 mm die with the aid of light steam. Immediatelyfollowing pelleting, diets were placed in a vertical drying towerto cool. After cooling, the pellets were screened and the remainderof the fish oil was sprayed (electric spray gun) onto the pelletsas they tumbled inside a cement mixer. The pellets were thenscreened (#10 U.S. screen- 2.0 mm) to remove any fines and thenstored in dark plastic bags in the cold storage room (2-5°C) forthe remainder of the experiment.5.2.3 Digestibility tanks and salt water filtrationThe 150 L digestibility tanks described by Hajen (1990) were118used with some modifications. A plexiglass cover was fitted overthe tank slit to allow for the passage of feces but prevented fishfrom entering the slit, which minimized scale loss. Water flowrates between 6-12 L/min were maintained to allow rapid transportof the fecal pellet from the tank to the settling column withminimal breakup of the fecal pellet. Feces expelled from the fishtravelled to the settling column in under 30 seconds.A filtration system was constructed for each of the 27digestibility tanks to prevent entry of sand, shells, plankton andlive organisms such as copepods which are pumped in with the saltwater. The system consisted of an inline Tee filter (AgriculturalProducts Inc.) with 0.150 mm screens which operated on thehydrocyclone principle. Incoming salt water was forced in acircular pattern around the screen which resulted in heavierparticles being displaced at the bottom of the filter and smallerparticles being trapped near the top of the screen. A ball valvelocated at the base of the filter allowed for flushing of theheavier sediment. The screens were cleaned and replaced when waterflow was hindered (once every two to three weeks). In combinationwith the inline Tee filter, bag filters (Knight Corp.-polypropylene-5 to 10 um mesh #1 size) were also installed toremove any remaining objects. The bag filters were cleaned bybackflushing with dechlorinated city water when required (every 24-48 hours).5.2.4 Aquarium facility and Experimental protocolChinook salmon juveniles (monosex females) were obtained from119a local hatchery (Seaspring Salmon Farm-Vancouver Island) and thenthey were transported to the West Vancouver Laboratory in May,1990. The fish were placed into stock tanks (3.6 m 3) supplied withwell water. One week before transport, the fish were handvaccinated to protect against vibriosis and furunculosis (Furogen-Vibrogen from AquaHealth Inc.). The fish were then acclimatedslowly to salt water over a 10 day period. Fish were fed 2-3 timesdaily to satiation with Biodiet. Since an unknown vibrio sp. wasdetected at heavy levels in 10% of the stock tank fish in August,the fish were fed feed medicated with Romet 30 (167 mg/kg fish) for10 days. The fish had grown to 15-25 g by August 24/90 at whichtime the fish were anesthetized (2-phenoxyethanol 0.5 ml/L water),selected for uniform size (15-25 g) and were then distributedrandomly into each of 24 tanks (40 fish /tank). The row of 24 tankswas subdivided into three blocks each consisting of 8 tanks. The 8test diets were randomly assigned to each of the tanks within thethree blocks (each diet was fed to triplicate groups of fish).The fish were fed the chromic oxide free basal diet for 4 daysat which time the fish were reweighed and fed the test diets forthe remainder of the experiment. After several days, fish in someof the tanks became bloated and they showed signs of dark skin, finerosion and tail loss due to tail nipping by other fish. Necropsiesperformed on dead fish revealed muscle degradation and haemorrhagesin the tail region while the spleen was swollen with soft tissuesurrounding it. The experiment was terminated on September 8 assome groups had reached 50% mortality and white fecal casts were120observed in the fecal collection tubes. At this stage, it wasdecided to run an experiment of the same design with rainbow troutin fresh water and then initiate another experiment with chinooksalmon when healthy fish became available.Rainbow trout were purchased from a local hatchery (SpringValley, Langley, B.C.) on October 9, 1990 and they were placed intostock tanks supplied with dechlorinated city water. Subsequently,they were fed basal diet twice daily to satiation for 7 days. Thefish were then anesthetized with Marinol (0.25 ppm in the tanks andtransport containers) and MS222 (70.0 ppm in buffered fresh water),weighed, selected for uniform size (56.0 g + 2 SD) and distributedrandomly to the 24 digestibility tanks with 35 fish/tank. They werefed basal diet (reference diet without chromic oxide) twice dailyfor an additional 5 days to allow for collection of chromic oxidefree feces. The rainbow trout were then anesthetized with Marinoland MS222 and re-weighed prior to feeding the test herring mealprotein diets. After this, fish were then fed the test diets twicedaily (three passes) to satiation (approximately 2-3% bodyweight/day). Daily records of feed consumption and mortality werekept and dissolved oxygen and water temperature were monitoredevery 4 minutes with an Aquaguard system. This consisted of 4oxygen probes, one temperature probe and a data logger integratedwith a PC computer to record and summarize data. Air flow ratethrough aeration tubing in each of the tanks was set to maintain atleast 80% oxygen saturation at the water outflow. The experimentalconditions are presented in Table 13.121Table 13. Experimental conditions during the rainbow troutand chinook salmon studies (Herring meal)Experiment no.1	2Experimental fish 	 Rainbow troutInitial mean wt (g) 	 56.14Final mean wt(g)	 77.58Number of fish/tank 	 35Stocking density (kg/m3 ) 	 13.10Number of groups/diet 	 3Type of tank 	 FiberglassWater volume (1) 	 150Source of water 	 City wateraWater temperature ( 0 Cb ) 	 9.3-11.2Water supply (1/min)	 6-8Salinity (ppt) 	 0Dissolved oxygen (mg/1)"	 8.5-10.0Feeding frequency/day 	 2Duration of feeding (days)	 20Photoperiodc 	 naturalStart time	 October 21/90Chinook salmon35.2939.70255.883Fiberglass150Salt water9.2-9.96-828-317.6-9.9218naturalNovember 21/90a Dechlorinated city waterb Monitored by AQUAGUARDc Photocell to Vitalite Durotest light122For the experiment with chinook salmon, juveniles (monosexfemales) were obtained from Seaspring hatchery and transported(with Marinol-Metomidate at .20 ppm) to the laboratory in October1990. The fish had previously been dip vaccinated against vibriosisand furunculosis (Furogen-Vibrogen from AquaHealth Inc.) prior totransport. The fish were placed into stock tanks supplied with wellwater and they were slowly acclimated to salt water over a 10 dayperiod.The fish were fed Biodiet twice daily to satiation until 4days prior to being transferred into the digestibility tanks. Atthis time the fish were fed the chromic oxide free basal diet (samediet as used for rainbow trout experiment). The fish had grown to25-45 g by November 15/90, at which time the fish were anesthetized(Marinol and MS222 combination), selected (25-45 g range) anddistributed randomly to the 24 digestibility tanks with 25fish/tank. The use of Marinol in the stock and transport tanksprior to capture and weighing of fish once again seemed to calm thechinook salmon as fish were easy to catch and were not skittishduring transport. The fish also recovered fairly quickly and beganfeeding the next day.The chinook were not re-weighed prior to feeding the testdiets to help minimize stress and prevent disease. Daily records offeed consumption and mortality were kept.All tanks were flushed and brushed (one third of tank drained)at the end of each day to remove any uneaten food particles andfecal material that may have accumulated in the tank and settling123column. Feces were collected in the morning from the settlingcolumn of each tank in 250 ml centrifuge bottles. The bottles werecentrifuged (10,000 g for 20 min.- Sorval RC5 RefrigeratedCentrifuge) as described by Cho et al. (1982; 1985) and Hajen(1990). The supernatant was discarded and the pellet was thenfrozen, pooled (daily collection for each tank) and freeze-driedfor further analyses.Fecal collection was begun three days after initial feeding ofthe chromic oxide test diets for both chinook salmon and R. troutto allow for complete passage of the chromic oxide through the fishdigestive system even though green feces was observed 36 to 48hours after feeding had commenced.Feces and feed were analyzed for proximate composition(moisture, ash, lipid and protein) and for the chromic oxideindicator concentration as determined by the wet ashingcolorimetric method of Stevenson and De Langen (1960).Fecal samples were freeze-dried and ground using a mortar andpestle. Since scale loss is a problem for chinook salmon in saltwater, fish scales (protein, keratin) contaminating the feces wereremoved using a .250 mm sieve as recommended by Hajen (1990).Screening of feces did not affect the chromic oxide values(Appendix Table 1). The rainbow trout feces were also freeze-driedbut because of the large volume collected, feces were ground in acoffee grinder (Braun). Coffee grinding did not affect the chromicoxide values (Appendix Table 1). Recoveries of chromic oxide fromCr2O3 spiked feed and feces are presented in Appendix Table 2.1245.2.5 Calculation of digestibility and fish performanceThe apparent digestibility coefficients were measuredaccurately by the methodology of Hajen (1990). The indirect methodfor determining apparent digestibility coefficient used in thisexperiment is based on the concentration of an indigestible markerin the feed and feces. The concentration of the marker is increasedwhen the nutrient is digested and absorbed.Feces collected from salt water are contaminated by salt. Someauthors have tried to wash the feces with fresh water to remove thesalt. However, this would result in leaching of organic nutrients.To overcome this problem, Hajen (1990) recommended thedetermination of organic matter digestibility instead of dry matterdigestibility. In this regard, the ratio between organic matterconstituents and the digestibility marker remains constantregardless of the level of salt contaminating the feces.At the end of the experiment fish were anesthetized andindividual fish lengths and weights were recorded. Specific growthrate was calculated as follows:SGR = in Final weight(g) - In Initial weight(g) x 100 duration of experiment (days)The feed efficiency ratio was expressed as:FE = wet weight gain (g) dry feed intake (g)The protein effiency ratio was calculated as:PER =  wet weight gain (a) protein intake (g)5.2.6 Pepsin and pH-stat determinationsPepsin digestibility was determined by the modified Torry125method (pepsin activity 1:10,000 at 0.0002% concentration) asdescribed by Lovern (1965) and modified by the InternationalAssociation of Fish Meal Manufacturers (Collaborative test methods1989). Briefly, 1.0 g of herring meal (finely ground) was mixedwith 150 ml of 0.0002% pepsin-HC1 solution (A) and placed into anagitator (Series 25 incubator/shaker) at 45°C for 16 hours. At thesame time, an acid insoluble determination was performed with 1.0g of herring meal and 150 ml of 0.075 N HC1 solution (B) also beingplaced in the agitator. After agitation the samples were filtered(Buchner filter fitted with Whatman #4) and the level of nitrogenwas determined on the filter and trapped sample. The pepsindigestibility was calculated as follows:% PD = % residual nitrogen B - % residual nitrogen A% residual nitrogen BThe pH-stat method employed in the present study measured thedegree of hydrolysis (DH) of the sample protein at 15°C for 500minutes by using digestive enzymes from the pyloric cecae ofrainbow trout with the aid of an auto-titrator system interfaced toa PC computer. Protein sources were analysed by Dr. Norm Haard atU.C.L.A. Davis. The method has been used to determine thedigestibility of a wide range of marine protein sources. (Haard,1991 WRAC Annual report).5.2.7 Statistical analysesComparison between digestibility values for the test proteins(7 sources) for each species was determined by one way ANOVA andfurther comparisons between species was performed with the student126t-test. Before performing analysis of variance to test forsignificant differences between treatments, Bartlett's test forhomogeneity of variance between treatments was performed asrecommended by Steel and Torrie (1980).The experiment was also analyzed as a 2 X 3 factorialexperiment with two processing temperatures and three levels of rawmaterial freshness. The control was freeze-dried herring meal.Correlation coefficients were determined between differentspoilage indicators over the storage and processing conditions.Correlation coefficients were also determined between spoilageindicator and digestibility values with salmonids and in vitrodigestibility values.5.3 RESULTS AND DISCUSSION5.3.1 Chemical composition of test ingredientsThe proximate compositions of the various herring press cakemeals are presented in Table 8. The protein levels ranged from 69.0to 74.7% of dry matter. Protein levels were lowest in the freeze-dried herring (FDH-7) but the lipid level was the highest (19.9%).The levels for protein, lipid and ash in the freeze-dried herringused in the present study are very similar to those reported forfreeze-dried herring presscake by Carpenter et al. (1962). Thehigher protein and lower lipid levels in the processed herringmeals are expected since heating and pressing the raw fish causesthe release of water and lipid which were not added back to thepress cake meals.The lipid levels were higher in the processed meals (12.7 to12716.2%) than levels normally found in commercial herring meal. Thelevels of lipid and protein did not seem to be affected markedly byspoilage or processing conditions employed. The levels of ashhowever, were highest in the meals produced from spoiled rawmaterial (LTSH-5, HTSH-6).The protein levels in the herring press cake meals wereslightly lower and the lipid levels were slightly higher than thosereported for Pacific whole herring meals by March et al. (1963) andAtlantic whole herring meals by Power et al. (1969). The proteinlevels compared closely but the lipid levels were higher than thosereported for herring type fish meal from U.K., Norway and Denmark(Barlow and Windsor, 1984; FAO, 1986).The ash levels in the spoiled meals (LTSH-5, HTSH-6) exceededthose reported for Pacific herring meal (March et al., 1963) andAtlantic herring meal (Power et al. 1969). The ash levels however,did not exceed the maximum ash content of 15% recommended by NRC(1981). Lassen et al. (1951) also reported an increase in ashcontent in fish meal as spoilage of the raw material progressed andthey attributed this to a loss of nitrogen due to microbialdegradation.The amino acid composition was affected slightly by processingtemperature since the levels of alanine, cystine, glutamic acid,glycine and particularly threonine decreased in the hightemperature meals (Table 9a,b). Spoilage of raw material did notaffect the level of amino acids with the exception of the mostspoiled meal (HTSH-6). The levels of alanine, arginine, aspartic128acid, glutamic acid, glycine and particularly proline and threoninewere lower in the spoiled meal.Lassen et al. (1949) found considerable destruction of aminoacids by bacterial putrefaction of raw fish that was processed intofish solubles. Losses of amino acids were very high for valine,arginine, histidine and leucine in the fish solubles produced fromhighly spoiled raw material. Toyama et al. (1983) confirmed thesefindings and reported that with putrefaction of the raw fish theamounts of total, individual and essential amino acids decreased inthe fish meal and soluble fraction. The amino acids greatlyaffected by putrefaction were aspartic acid, proline, arginine,alanine and lysine. However, Halland et al. (1988) reported littlechange in the amino acid composition of herring stored at 2°C forup to 21 days, with the exception of histidine and lysine whichdecreased during storage. Halland et al. (1990) also found littlechange in the amino acid content of mackerel stored at 2°C for upto 10 days and they concluded that reductions in lysine coincidedwith increases in cadaverine level, especially in whole fish. Thus,only during advanced spoilage of fish is the amino acid compositionaffected.Sandfeld (1983) noted that levels of amino acids decreasedslightly in fish meals made from previously spoiled raw materialbut he concluded that overheating (> 100°C) during fish mealproduction greatly decreases the levels of thermo-labile aminoacids such as arginine, cystine, lysine, methionine, serine andthreonine. Such destruction, however, was not apparent in the129present herring press cake meals.Most of the herring protein sources met the essential aminoacid requirements of rainbow trout, with the exception of cystine,methionine and tryptophan which were low in all meals. Similarly,most of the essential amino acid requirements for chinook salmonwere met, with the exception of histidine (low FDH-7), methionineand phenylalanine which were low in all meals.The mineral composition varied between the test herring meals(Table 10). The levels of calcium, increased as spoilage of the rawmaterial progressed while the levels of magnesium, manganese,phosphorous, zinc and potassium decreased as spoilage progressed.It is not known whether the minerals were lost in the press liquoras the amount of press liquor produced during processing alsoincreased with spoilage of the raw material (Table 14). The levelsof sodium were highest in the meals produced form spoiled rawmaterial (LTSH-5, HTSH-6).The levels of chromium, iron and manganese were alsoinfluenced by processing temperature because levels were lower inhigh temperature meals in comparison to low temperature meals(Table 10). Since temperature of processing also affects the amountof press liquor released during processing, the high temperatureresulted in the release of larger amounts of press liquor but it isnot known if the minerals were lost in the press liquor.The levels of calcium, copper, manganese, phosphorus, zinc andespecially iron were higher in the processed fish meals than thefreeze-dried herring. Flaking of rusted material from the lining of130the drum dryer during meal production may explain the high levelsof some of the minerals, most notably iron.The levels of calcium, iron, magnesium, phosphorus and sodiumwere higher for all processed meals in the present study comparedto the mineral levels reported by Power et al. (1969) for Pacificherring meal and for herring type meals (Barlow and Windsor, 1984)(Table 2). The levels of minerals do not exceed toxicity levels forsalmonids (Table 2) but high mineral contents are known to increaseoxidation of fish lipids during storage (Hsieh and Kinsella, 1989).The yields of fish meal and press liquor were influenced byprocessing conditions and spoilage of raw material (Table 14). Asspoilage progressed in the raw material, the total yield of fishmeal decreased and the level of press liquor (oil, solids andwater) increased. Lassen et al. (1951) also reported a decrease infish meal yields and an increase in press water and solids asspoilage of the raw material progressed. They concluded thatspoilage of raw material resulted in a 23% reduction in total fishmeal yield and the resulting condensed solubles from the pressliquor were of poor quality. Lassen et al. (1949) reported thatcondensed fish solubles produced from spoiled raw materialdecreased the growth of chickens in contrast to solubles producedfrom fresh raw material.The higher processing temperatures employed during mealproduction resulted in an increased yield of press liquor from theraw material. Ward et al. (1977) also found an increase in theamount of press liquor and solids released from both sand eels andI	 I••••••■0 v. 0 01 Ne --s1/40• • 	 • 	 •1-4 cy CO 0 COe-♦ e 1 	 1-1..••••CO 	 kf, CO %.0 CO CON 	 N 11• • 	 • 	 • 	 • 	 • 	 • 	 • 	 • 	 • 	 •LO 0 1-1 01 I-1 1/40 1/40 0 01 0 r"N M 	•:1'	 •4' 	 CI 	 •v:P 1‘) 	 lf1‘•-■ 	 1/40 	 cl)cr■	In	 szr 	 O 	 N 	 co1/40 	 4‘) 	 in 	 to 	 -zr 	 -zt•1-1 	 1-1 CO 01 01 VD 0 CO 0 kO N N• • 	 • 	 • 	 • 	 • 	 • 	 • 	 • 	 • 	 • 	 • 	 •	cNi 0 1/40 1/410 •ct' 	 M N M N In M CO 1/4.0N H 	 1-4 t-I 1-1 	 N N•••••• 	 1/40 	 0101 co 	 „ .1 01 N •;1' 0 O •:1' 1/40 0 	 . C71• . 	 • • 	 • • 	 • • 	 • • 	N •V) in .1' r-1 CO 1/40 el N COOL %.0 01o 	 in inco c.,1 	 en co 04 s N 01 N 01 N 	 •:1'1-1 	 1-1	.....- 	 ...... 	 ....... 	 ..-.•-......1-1 	 N 	 M 	 -zr 	 In 	 1/40 —	I—. I—. I 	 lino x o x co x co x N x - • I 0Nv 	zz 	 CI) 	 Cr) t-N)E-1 	 E-1 	 E-1 	 E-1 	 (-1 r-1)4131•vra) —4.) 1-4a)'b0oOH oEna)C.)O04 	 o 3o	 to,4.)14 4 jo 	 VI 'Cm• w4.) 	 a) wr1 A(c) 	 ).4pa) ' ,`1.1 0 a)(0 a) $.4W0:1 	 00 $4 	 El 4 0N C.)••-i01 4-4-r-1 0 	 1:4r-14-4LI)a) 4)(4$4 0 	 0 4 044-) 	 trilX'cl 	 u) a)4-111:1 	 44r-1qll 	 (11 Co-)r-1 	 'CS Inw N4j 0-r-1O ro0U)cUa)U)4.) 	 a)4.)4-1N	 •r--111:1r-1	$-1 	 $-1a)4.)	0 	 $-4•--1 	 (1-Io (13a) 	 a). 0• Wa)	4-)1:5 	4	44 	 Li)4-4 ni0 4 3gU)O N 	 U) 3 	 .1•14-1 En 	 a)0 	$4 tri	 4-)• 10.4ri 	0 • cri0'1 	w	 (4 0C --1 	4)	 ca0 X 	 M a)'H 	 0 a) 41 >44-) E 	 4-)0 $.4 rri7:1 	 0 a)rt:$ 	 C.) 4-) 4-4al (U 	0	 4-) 0cl)O 	 4 0 a)	I:1 4.) 	 T.)4-4 	 a) 	 >4Ts 	 oCO 	 4-)-1	0	 •rti 	 rCI 0 	 4	4.) w 	 o wrrs 	 o 	 g, 	 TS 0 	 4-1•zi,	"7.- 	 0 	 04-) 	 r0 	 3w 4-1 ;` 	 u) a) '0 0W 	 - 3 0 u) a) $.4„- 	 ,10 4 L- 0 $4 	 0E-1 	 a) a)• 4-)	 4)o (15 	 (11(Ci 	a) z'CI a)a) 0 	 u) 0 	 4-14.4 4-) 	 W 0	 0 0WO 	 .4f-10E-i 	 c0r) 01/40132sprats processed at elevated temperatures.The highest yield for fish meal was found for the freeze-driedherring since only water was lost during processing in contrast tothe herring press cake meals where water, lipid and some solidswere not added back during processing.5.3.2 Spoilage indicator levelsThe spoilage of the raw herring during storage and the levelsof chemical indicators in the raw product are discussed in chapters3 and 4. The influence of processing of the raw material onchemical indicator levels is also discussed in chapter 4. Briefly,the meals produced from the most spoiled raw herring had higherlevels of putrescine, cadaverine, and total volatile nitrogen. Thecommercial herring meal was also high in TVN and trimethylaminelevels. Cadaverine appeared to be the most useful indicator becauselevels were very high in the spoiled meals in contrast to all otherindicators (Table 11).The peroxide values (PV) (lipid quality) were highest in theherring meals produced from fresh raw materials but it is unknownif PV levels were increasing or decreasing in the meals since PVlevels initially increase then decrease during lipid oxidation. Thelevels of thiobarbituric acid (TBA)(u moles of malondialdehyde)were not monitored in the herring meals for several reasons.Erratic results were obtained with the Jumper meals used in thelast experiment (chapter 6). Also, losses of malondialdehyde duringdiet preparation and steam pelleting have been reported (Hung etal., 1981) and interferences from copper and iron (present in high133levels for all herring meals used in the present study) have beenreported with the TBA method (Hoyland and Taylor, 1991).5.3.3 Fish performance and apparent digestibility valuesThe specific growth rates, feed efficiency ratio (FE) andprotein efficiency ratio (PER) values in rainbow trout appeared tobe unaffected by spoilage and processing conditions of the herringmeals (Table 15). The growth data were not statistically analyzedsince diets were not designed to be isonitrogenous or isocaloricfor growth studies (30% of the diet was the test ingredient) andthe experiments were only of short duration. The feed efficiencyand PER data were not reported for chinook salmon because they werefed to excess to obtain an adequate feeding response. This resultedin some wastage of food which adversely affected the FE and PERvalues. However, both the chinook salmon and rainbow trout were ina state of positive nitrogen balance as demonstrated by the growthdata (Tables 13 and 15).The apparent gross energy and protein digestibility values forthe test herring meals in chinook salmon were not significantly (P>0.05) influenced by the higher drying temperature or the degree ofspoilage of the raw material (Table 16). By contrast, proteindigestibility was lowered by advanced spoilage of herring inrainbow trout similar to chinook, thermal processing conditions didnot influence protein digestibility in trout. Other studies haveshown decreased protein digestibility in salmonids, mink, andpoultry fed diets containing high temperature (> 100-140°C) fishmeals (Tarr, 1982; Sandfeld, 1983; Opstvedt et al., 1984; Pike et•rIa)z orx.ia, 04Cra)a)4444a)cla-r-109•44Vaa)-04-)0>1134>, I-1Cl) 	 1::$•r-I 	 0C.) 01 C.)444-1 $.4 C..)W $4 )4W4 ow4-) 4 g0 4.)8.10a) ;..44-4Itp 	 00I-I- 	 044 U) ••••1WTs 0 ga) zar.4 4444 	 a)OP 'CI(1) r) $.4-6) `"••• 0(U)4 0'1g Co.0 .1-14-) g a)•-1o$-1 4-) ftU o4'4 rn• 4-)u 	 a)(4-1mai ro a)4J a) •444 tn•r1 -6)w 0 '03 0 0r-1 4-1 CV$-1•r4 0 044 A 44'CI 	 u)(0 a)oo ou)4.)Ea gW-6)(1:1 0$-1a) 0A 4) u)E-1 $44-)co r4 cf) r) U1 CV VD VDcr 10 .4. to in If) 	 II/• • 	 • 	 • 	 • 	 • 	 • 	 •0 0 0 0 0 0 0 0•:1' 0 N CO N 	 CO CVCO 01 CO 01 CT 01 01 01• • 	 • 	 • 	 • 	 • 	 • 	 •0 0 0 0 0 0 0 0In CT VD CO CV CO I-1 CVl0 VD N N l0 N N• • 	 • 	 • 	 • 	 • 	 • 	 •0 0 0 0 0 0 0 00 N 1C) 01 01 1-1• • 	 • 	 • 	 • 	 •	Cl 	 VD 	 01 	 r-I 	 CV 	 1•1	r-I 0 01 	 N O1 CV	CO CO N CO 	 N CO CO	CU 	 1•1 CN M Tr in ■0W	W 	COcI	WOO 	 44 44 Z Z ci) cr) x 44E.-4 	 E- 	 0 	 41	ra4 Er1	 a x 14 x 14 x w 1=4a)a)a)0(U00'10>4 	 a)4-4■■•• 01-1 13)II3 C.)4-) W4-) a)04-) $-4 .4-111 4-1>44-4u) 4-) 	 W34-1 0 W)-1in tT1 U a)Cl -.4 4-144 04-i • r 14-104-1 U-'-4C U I:I II(r1 a) Wa) 	 (1)C/3 C7.44 .0 U 9CO Cl CO CV 0 0 CO CO	In r-I 	 CO 01 N• • 	 • 	 • 	 • 	 • 	 •If) 1/40 LC) CO N le) CV NNNNNNNCON	CV 	 ri 	 CV 	 lf) 	 V.) 	 IC) 	 1.0OCIONCNI-ICON• • 	 • 	 • 	 • 	 •	VD I.CIVD•zr 	 ISIVD01In In in to to WI 10E-1acc)0cif,E-4aU)0a)4-)4-)faU--4=(Utr0-.g4 a)a) 04-) 4.-) $4W0 0•1-1 $4 00 Ia• cnE-Iacc)0-2 	 20 CO• •N. OCO CO2sze Cr 01• • 	 LOtN COCO N I-I135al 4O W-ri a))4 P$.4 44a).g Z-.444O 4.)0••-'s 013:1Z 4.)0.....0Ul 4W z0 -?-41--1 ms(0 4'CIt:r1 (13$4(11 ..•...Ci)CD 0......W41 a)-ri 4 j z4.3m 0 0N 3 -,-1.4' 4) 4j0:iIti 71:1/ 20U),O	 44O• g g-.4 -1-1Eil•:11 (.9 4T-14.) 0 Z,--1 0- ,-1 .r.1r.-/ (13-,-.1 cn :,t.1:2 	 ro,-I .g0 7:4N a) c )4„t:411. .zr.1 0'Ti'.' (444.) 0 CoZ 4-' (1,)a) '.'I-.4 'T3 ...q• Cli (04-) 4-1 E.40Z U) 0CU 4.)-P 0g $4 ..ia) 	 a)$4 0 4.4al ca0 (1)P•Cli C Z •••••4 • ria) .....• 4_, E..,m) 0 c4.4 $.4 •••••C-14O )-I0-1 4-) a)A W 4-)0 (1) asE4 4CWWWW 	 *C C C 44 44 *.14A An A Aa,	 i	 in	 1/4c,	 *ON Z.0 -1 	 tr,• •	 •	 •	 •:t.O 0 e4 0 CO CO (NtN 01 IN CV .-I 14 O3 0-2 	 -2 	 -2 	 .0e1 	 1-1 tio 'clY1 42E■ 	 If) 	 N 	 tv)• • 	 • 	 • 	 • 	 • 	 Lo1-1	 v-I .4 	 ,-1 a,	 c, crCN 	 CV 	 CN 	 CN f-4 (.1 I.1 0ICW W w W -IC * 4cO 0 C gin d'•:). ;• 	an 1.	 2	m:) 	 1-4 	 4	 e•-,	 N.	 •	 •	 •	 •	 cr,O1	 co	 c,	 N	 •:t.	 .-4	 ,-.4	 •co co a, Co co co 01 I-4Z0 01 731 °c:,• • 	 • 	 •0 0 1-4 (,)01 O 01 01	731 'IL JZO 	 e-I• • 	 •	 r.N al 1-ico co co CVEn	 cn _u)	 EnO z C z -r.-2 	 .2 	 .1 	 2 	 2.--i in Ct■ 0 U	 CV C*1 01• • 	 • 	 • 	 li)01 0 N CO '.0 N I"-	 •CO CA CO CO CO CO CO 07-4 	 11-Co 	 7-1 	 Zo 	 11., A,-.1 	 01. 	 • 	 c•-,a,	 Ol 	 1-1 	 .4 1-1 (N el	•CO 01 01 01 01 01 co r)..-,ca ul ca WWW*C g C CC0CH C NI eb C  V 0 . " 1/4 D C r .i.• •	 •	 •	 •	 •	 •	 ,1E",	'.0	 '.0 	 LO 	 01 	 N 	 If)CO CO CO CO N Is- CO 14,7.o 	 7--) IIN•N r, Li)CO CO COv-I 	 CV 	 C•1 	 'cl'	If)	 ,c)IIIIIIhmxx xx= I44 44 Z Z En CI) = ZEl Ei E+ E-i E4 E4 0 411-4 = 1-1 = 1-4 = 44 cn.- 404-) 	 0O IIa) 	 C14a) 	 4:44 	 le44-4 	 'LO'Ti 	 0>.• 	 c;•-I4-) 	 IIC 04O *C.) 	 ..-- ,-144 	 .I..)-.4 	 Eng W01 	4..).1-1 	 Ica 	 4.)4-) 	 rrio a)O S-1..-4a) 	 0W 04O 454 	 4..)W 	 "-I4-) 	 34-1 	rt:$a) 	 a)ES34-1 	0.)$:14 	 4)• .4-1CVI $4 04O U) 	 cll-r-I P 	 34.) W0 04 	 Eli_O 0	a).--i W 	 -r-I0S.4• 0 	 El)a) 0 	 04-P 0	 EnW 0.r) 0 	 gU Q)a) 	 a)a) ni 4.)4 	 a)4.) 4.)	 4--)3 	($)(4-4 	 wO ----- 0g In gC 0 O w(0 0 •CU r-I 0 wO 0 VI (4.4O al 4.4W 	 -44 4 -1-1 ro •4..) 0 u) 	 ...-.(U W ■-4 e--1a) a) 4-) ra 054 	 0 0(U 0 W -,-1 •-4 •• 4.J 0Ell >I En IIa) Ea a) -.4 04O g 	 4-)r-I fa 0 it *cts a) El 4-) *O Z •--- cn IC* 	 EnO136al., 1990). The reduction in digestibility of the high temperaturemeals is thought to result at least in part from the increasednumber of disulphide bonds formed during excessive heating of rawfish (Opstvedt et al., 1984).Njaa et al. (1980) reported that spoiled raw material(oxidized lipid) had no effect on protein digestibility of the dietfor chickens, rats, and minks. Watanabe et al. (1983) and Austrengand Gjefsen (1981) confirmed that spoilage of the raw material(lipid oxidation) did not affect the protein digestibility of thediet in rainbow trout and Atlantic salmon. Although, Pike et al.(1990) found that fish meals produced from spoiled raw material atlow temperatures (60°C) decreased their protein digestibility inmink but staleness of raw fish did not affect protein digestibilityin high temperature meals (140°) in minks. The findings of thisstudy with respect to trout agree with those of Pike et al. (1990).Organic matter (OM) and digestible energy (DE) coefficientswere not significantly (P<0.05) affected by the higher dryingtemperatures but the values were significantly (P<0.05) depressedby advanced spoilage of the raw material in both chinook salmon(organic matter only) and rainbow trout (Table 16). Similarly,gross energy contents of the test protein sources weresignificantly depressed (P<0.05) by advanced spoilage of the rawmaterial in rainbow trout.Digestibility values for, crude protein and energyof the processed fish meals were generally higher in chinook salmonthan in rainbow trout. The opposite occurred for freeze-dried137herring where all components were better digestible by rainbow thanby chinook salmon. Digestibility values (OM, CP and GE) weresignificantly different (P<0.05) between species for the freeze-dried herring meal and the meals produced from poor quality spoiledraw material.The low digestibility values for the freeze-dried herring inchinook as compared to rainbow trout were not expected sincefreeze-dried fish have been reported to be well digested by rainbowtrout (Opstvedt et al., 1984) and plaice (Cowey et al., 1971;1974). However, reductions in available lysine, in vitro enzymaticdigestibility and overall digestibility of the whole diet insalmonids have been reported for freeze-dried fish protein sources(Fosbol, 1985; Koizumi et al., 1990).The low moisture content of freeze-dried fish allows forincreased lipid oxidation and browning due to lipid and Maillardreactions during storage (Tarr and Gould, 1965; Gardner, 1979;Bligh et al., 1988). Therefore, lipid oxidation of the freeze-driedherring may have occurred since antioxidants were added to theherring only after freeze-drying and grinding. Also, Austreng andGjefsen (1981) noted that Atlantic salmon are particularlysensitive to quality of dietary fat when compared to rainbow trout.Chinook salmon may also be more sensitive than rainbow trout toquality of dietary lipid.The chinook salmon in the present study may have been verysusceptible to quality of dietary lipid since the fish had recentlybeen transferred to salt water. It is well known that transfer of138salmonids (chinook and coho salmon) into sea water causes depletionof lipid reserves due to hormonal and possibly osmoregulatoryeffects (Sheridan, 1988a; 1989). Hence, it would seem logical thatchinook salmon recently transferred to sea water would be moresusceptible to the quality of dietary lipid than rainbow trout infresh water.The higher protein digestibility values for the test proteinsources in chinook salmon, even for the meals produced from spoiledraw material in comparison to rainbow trout may have been theresult of species differences and/or the influence of salinity orother culture conditions.Chum salmon have two pepsins (I, II) located in their gut andonly when the fish are in salt water are both pepsins active. It isnot known if a similar isoenzyme system exists for chinook salmon.Lall (1988 as cited in Higgs et al., 1990) found that Atlanticsalmon in sea water had higher protein digestibility coefficientsfor several feedstuffs than Atlantic salmon in fresh water. Incontrast, a decrease in apparent protein digestibility withincreasing salinity has been observed in rainbow trout (MacLeod,1977, Lall and Bishop, 1979) Atlantic salmon (Usher et al., 1990)and arctic charr (Ringo, 1991). Salinity has been shown to increasethe gastro-intestinal rate of food passage in salmonids (MacLeod,1977; Collie, 1985) and to increase the demand for free amino acidsin the gut for osmoregulation (Jurss et a/. 1983; Dabrowski 1986)which may account for the decrease in protein digestibility.The decrease in organic matter digestibility values of the139meals produced from spoiled raw material in both chinook salmon andrainbow trout may reflect a reduction in lipid digestibility. Sincelipid is very high in energy, the digestible energy values werealso adversely affected by spoilage of the raw material. Oxidationof dietary lipids fed to chickens (March et al., 1965; Opstvedt,1973a,b), rainbow trout (Watanabe et al., 1983), Atlantic salmon(Austreng and Gjefsen, 1981) and carp (Iijima et al., 1983) hasresulted in a reduction in the digestibility of the lipids.Watanabe et al. (1983) concluded that lipid oxidation in brown fishmeals was the most important factor in reducing the overalldigestibility of fish meals in rainbow trout.Oxidized lipid has been reported to diminish the activity orinactivate trypsin, pepsin, alpha-chymotrypsin and succinatedehydrogenase (Roubal and Tappell, 1966; Hata and Kanida, 1980).Oxidized lipids are highly reactive and they may destroy labilevitamins and reduce the availability of some essential amino acids(Carpenter et al., 1963; El-Lakany and March, 1974; Gardner, 1979;Opstevdt, 1985).In general, the protein digestibility of herring meal byrainbow trout ranges from 73.2 to 87.5% (Asgaard, 1988) but valuesup to 92% have been reported (Cho and Kaushik, 1990). The proteindigestibility values for the herring meal in the present study forrainbow trout and chinook salmon were very similar to thosereported in the literature for commercial herring meal in rainbowtrout (Smith et al., 1980; Cho et al., 1982), chinook salmon(Hajen, 1990), carp (Hossain and Jauncey, 1989) and sea bass140(Spyridakis et al., 1988a). Similarly, the organic matter andenergy digestibility values obtained in the present study withherring meal in chinook salmon and rainbow trout compare closely tothose reported by Hajen (1990) and by Cho and Kaushik (1990),respectively. Digestible energy values for the meals produced fromfresh and medium spoiled raw herring were higher than DE values formeals produced from spoiled raw material in both species. DE valuesfor the meals from spoiled raw material were lower than reported DEvalues for commercial herring meals in chinook salmon (Hajen, 1990)and rainbow trout (Smith et al., 1980; Cho and Kaushik, 1990).5.3.4 Pepsin solubility and pH-stat valuesPepsin solubility values were low in the herring mealsproduced from the fresh raw material (LTFH-1, HTFH-2 and FDH-3) andthey increased as the spoilage of the raw material increased.However, the increased processing temperature for the meals hadlittle effect on pepsin solubility (slight decrease in hightemperature meals) (Table 17). A possible explanation for the abovefinding may relate to the extent of enzymatic degradation ofproteins by enzymes produced from bacteria and from within the fish(autolysis). Pike et al. (1990) also reported that pepsin valuesincreased in the meals produced from spoiled raw material and thata higher processing temperature of the meal decreased pepsinsolubility only in the meals processed from fresh raw material.Thus poor quality meals produced from spoiled raw material wouldgive a misleading estimate of protein quality in this case.The pepsin method has been criticized for lack of sensitivity141Table 17. In vitro digestibility of herring meal protein sources.(Values are the mean of two determinations and are expressed on adry matter basis. Refer to table 8 for additional information.)Protein Source Protein Digestibility%Pepsin Solubility'pH-stat Digestibility 2Degree Hydrolysis %LTFH-1 79.6143 18.28abHTFH-2 77.99d 17.19bLTMH-3 80.40e4 22.08aHTMH-4 81.90' 21.73aLTSH-5 88.64a 18.6931'HTSH-6 86.86b 21.65aFDH-7 49.404 	 (68.33)' 19.10 abSEM 1.72 0.56'Determined with 0.0002% pepsin with 1:10000 activity.2 Determined with enzyme extract from cecae of rainbow trout.3 Means in each column with a common superscript letter are notsignificantly different (Tukey's test P>0.05).4 Lipid extracted meal in parentheses (chloroform/methanol).142when comparing good quality and excellent quality fish meals (Hardyand Masumoto, 1990). Nevertheless, the test is able to distinguishbetween the quality of overheated(>100°C) and low temperature fishmeals (Lovern, 1965; March and Hickling, 1982).Pepsin solubility was lowest for the freeze-dried herring,even when the lipid was extracted. It is possible that theremaining lipid in the FDH lipid extracted meal may have still beeninterfering with pepsin, since a high lipid content is known tointerfere with pepsin solubility (March and Hickling, 1982).The pepsin solubility values for the processed herring mealsin the present study were much lower than those reported forAtlantic herring meals with 0.2% pepsin (93.9 to 95.4%) and .002%pepsin (90.0 to 91.7) (Power et al., 1969) and for Pacific herringmeal with 0.2% pepsin (92.1 to 93.0%) (March et al., 1963). Boththe high lipid levels of the herring meals and low pepsinconcentration (0.0002%) used in the present study may havecontributed to the low pepsin solubility values.Pepsin digestibility values for the test protein sourcescorrelated the best with gross energy digestibility values inrainbow trout and with protein digestibility values in chinooksalmon (Table 18). Otherwise pepsin solubility correlated poorlywith salmonid digestibility values.The pH-stat values for the test protein sources (Table 17)also correlated poorly with salmonid digestibility values (Table18). PH-stat values were only slightly influenced by higher mealprocessing temperatures in the fresh and medium spoiled raw143material but spoilage of raw material did not seem to affect pH-stat values.The level of the spoilage indicators, putrescine andcadaverine, in the herring test protein sources correlated thehighest with organic matter digestibility values in rainbow trout.However, the correlations between indicator levels and chinookdigestibility values were poor (Table 18). Thus, it would seem thatthe level of one indicator alone can not be used to predict thenutritive value of herring press cake meals. Cadaverine levelshowever, show the most promise in predicting the nutritive value ofsuch meals in the present study.5.4 CONCLUSIONSHerring press cake meals prepared from fresh to moderatelyspoiled (held for up to 8 days at 2-5°C) material were verydigestible for both chinook salmon in sea water and rainbow troutin fresh water. Meals produced from spoiled raw material (held for12 days at 2-5°C) had an adverse effect on organic matterdigestibility and digestible energy values in chinook salmon and anadverse affect on all digestibility values in rainbow trout.Elevation of processing temperature from 75°C to 100°C did notadversely affect digestibility values in rainbow trout or chinooksalmon. Protein digestibility values were affected the least byspoilage and processing conditions of the raw material in bothspecies.Digestibility values for the freeze-dried herring and herringmeals produced from spoiled raw material were significantly144Table 18. Correlations between various digestibility values andspoilage indicators of Herring test protein sources for pH-stat,pepsin solubility, organic matter, protein and gross energydigestibility with rainbow trout and chinook salmon.Comparison 	 R 	 R2IN VITRO VS IN VITROpH-stat vs pepsinIN VITRO VS FISHpH-stat vs OM-RT 1vs 0M-CHINvs NI-RTvs NI-CHINvs GE-RTvs GE-CHININ VITRO VS FISHpepsin vs OM-RTvs OM-CHINvs NI-RTvs NI-CHINvs GE-RTvs GE-CHINFISH VS FISHRT-OM vs CHIN-OMRT-NI vs CHIN-NIRT-GE vs CHIN-GE.278 .077- 	 .240 .058- 	 .025 .001- 	 .406 .165.206 .042- 	 .202 .041.332 .110- 	 .690 .476.173 .030- 	 .352 .124.807 .651- 	 .818 .669.561 .315.519 .269.167 .028- 	 .179 .032INDICATOR VS FISHOM-RT NI-RT GE-RT OM-CHIN NI-CHIN GE-CHINPUT -.980 -.638 -.931 -.453 .339 -.107CAD -.988 -.682 -.925 -.465 .367 -.088HIST -.454 -.488 -.217 -.857 -.586 -.823TVN -.923 -.767 -.744 -.686 .191 -.329TMA -.697 -.123 -.891 .185 .715 .496PV .760 .316 .834 -.033 -.843 -.449INDICATOR VS INDICATORPUTPUT 	 1.0CAD HIST TVN TMACAD .993 1.0HIST .388 .358 1.0TVN .894 .924 .473 1.0TMA .708 .685 -.001 .405 1.0PV -.753 -.784 .154 -.674 -.794RT= Rainbow trout, CHIN= Chinook, OM= Organic matter, NI= Nitrogen, GE= Grossenergy, PUT= Putrescine, CAD= Cadaverine, HIST= Histamine, PV= Peroxide value,TVN= Total volatile nitrogen, TMA= Trimethylamine145different between species. The lower digestibility values for thefreeze-dried herring meal in chinook salmon in salt water may haveresulted from lipid oxidation. The greater adverse effect of rawmaterial spoilage on organic matter digestibility than on proteindigestibility, would suggest that lipid digestibility would havebeen adversly affected by spoilage. Lipid digestibility values werenot determined in the present study.The in vitro digestibility values (pepsin and pH-stat) werealso influenced by spoilage of the raw material. However, theincreased values associated with progressive spoilage of the rawmaterial indicates that these tests would over-estimate in vivodigestibility values of these meals. Processing temperaturedecreased the pepsin solubility values slightly in the hightemperature meals.146Chapter 66.0 EXPERIMENT 4. Influence of raw material freshness and ofprocessing conditions on the nutritive value of ocean perchprotein hydrolyzates (Jumper Feed Systems Ltd.) for rainbowtrout in fresh water and chinook salmon in sea water.6.1 INTRODUCTIONOcean perch (Sebastes alutus) is a typical underutilizedgroundfish species that has been harvested in increased amountsover the years. For instance, over 6,165 tonnes were exported fromCanada and 10,130 tonnes were landed in U.S.A. in 1989 (Dept. ofFisheries and Oceans, 1990). The sustainable yield for ocean perchoff coastal British Columbia was set at 3,350-5,470 tonnes/year for1990 (Tyler and Fargo, 1990). With such large amounts of thismaterial harvested or potentially available, an investigation of amethod for converting this species into a protein hydrolyzate wasbegun by Jumper Feed Systems Ltd. with equipment manufactured byAdvanced Hydrolyzing Systems Inc.A protein hydrolyzate is produced in a fashion similar to thatused for type B fish protein concentrate (FPC). The term FPCusually refers to fish powder (meal) intended for human consumptionthat has been produced from raw, wholesome fish under hygienicconditions similar to those processes used in the production offish meal (FAO, 1986).In contrast to FPC, a protein hydrolyzate utilizes proteolyticenzymes to help extract proteins and lipids during subsequentthermal processing and evaporation.The enzymatic process is similar to silage production sincethe raw material is ground, digested, pasteurized, screened (to147remove bones and debris) and decanted to remove water and oil.Hydrolyzates are then thermally dried (evaporators) to less than10% moisture and stabilized with antioxidant. The proteolyticenzymes used in the production of protein hydrolyzates may beeither of vegetable or microbial origin, thus differing from fishsilage where enzymes that are naturally present in the fish fleshare used for liquefacation. Fish silages are usually not thermallydried since they incorporated into moist feeds (>30% moisture). Theproduction of fish silages has been reviewed by Tatterson andWindsor (1974), Raa and Gildberg (1982) and Jangaard (1987, 1991).Fish hydrolyzate, is a light brown powder containing highlevels of protein and lipid (up to 35% in powder form) and lowlevels of ash. The hydrolyzate is thermally processed attemperatures of 40 to 60°C which are cooler than temperaturesemployed during fish meal production. The enzymatic process isconsidered to be gentler on the protein than are processesassociated with fish meal production. Also, hydrolyzed proteinsretain important functional properties such as dispersability,solubility in water and emulsifying properties (Windsor and Barlow,1981). The production of fish hydrolyzates has been reviewed byWindsor (1969), Hale (1972), Sikorski and Naczk (1981) and Windsorand Barlow (1981).It has been shown that processing temperature and the natureof the raw material influence the quality of marine fish proteinand therefore protein digestibility and salmonid performance (Tarr,1982; Opstvedt et al., 1984; McCallum and Higgs, 1989; Pike et a/.148, 1990).To establish the optimal processing parameters of thehydrolysis equipment, a two by three factorial experiment wasdesigned with two drying temperatures (85 and 93.3°C) and threelevels of raw material freshness. These meals were tested byconducting growth and digestibility trials with rainbow trout(Oncorhyncus mykiss) in fresh water and chinook salmon (Oncorhyncustshawytscha) in salt water. The levels of TVN, TMA, PV, and TBAwere measured in the meals immediately after preparation. Pepsindigestibility (modified Torry method - very dilute) and the pH-statmultienzyme digestibility values were also determined for the testprotein sources.6.2 MATERIALS AND METHODS6.2.1 	 Test protein sourcesThree hundred kilograms of ocean perch were prepared by JumperFeed Systems using Advanced Hydrolyzing Systems Inc. equipment.The end product was produced by autolytic enzymes and thermaldrying followed by screening to remove excess bones and debris.The process is similar to that used for other fish proteinconcentrates (Hale, 1972; Stone and Hardy, 1986; Hardy, 1987) butthe pH is not adjusted with expensive chemicals. Also, costly spraydrying is not used. Prior to processing, the fish were stored for0, 4, and 8 days at 2-5°C, with each batch then processed at 85°Cand 93°C. A herring meal prepared from fish stored for 8 days andprocessed at 77-95°C was used as a control. The six test proteinsources and reference protein source are described in Table 19.coaQ)2.=OcoCa)0Lo_-4-1MI-ILO1COI--.1CViI-I011MI--—.1149o 	 o 	 •:t. 	 •o- 	 co 	 co 	 coLOto 	 co	 to 	 co 	 to 	 co 	 a)Co	 a) 	 co 	 a) 	 co 	 a) 	 IN.N-co 	 N-I 	 Ici) 	 II— 	 a= 	 cc.c 	 -C0 	 08 	 8	 .....,73 	 73 	 -c) CI 73 CL 13 	 73 	 e-a) a)a) 	 a)._	 a) L. 	 •_ c .r: C .- .0 1-..- _c toL.. co	 -10 (CI 	 )... c.) 	 1373 .0 73 -C 73 0	CI) 13 /- 	 2 co... 	0 	 W CO .71L_G) Oe 2 2 	0Q) 2 0 	0	2  0 	 0.L_ 	 2 a =a = a = 	 = 	 = c = 	 —i 	E— c 4 	 c4-t; 	as _c	c 1.- c " co to c	 c' _	co 	 i-' 's 	co coL. a) L_ 0 .773 c)N co a) co 	 0 a) a) a) 0 0 CD 0 	 ._a (1) CL 0 a.!...- a.:_.- a0 a 0 CD "E 	 c.) 8° E °O 0 a) 0 E 	 E 	 Ea) E .2 E	 a) 1z3 	 cp -10 	cDC C4-. = 	 M 4-. 0 	a)	 L_ __c 	 _c 	 :- .0 = a) co..c co	'5 -c '6 	 3 	 cn cp	0 w	O 2 2 w 	a)f■ • — 1,-. 	 o 	 ._o o. — a a)- J 	 2 4- ..., E i E -J (/) = a) CC150The proximate composition and gross energy contents (Table 20)of the test meals were determined as follows; moisture (20 hrs at85 °C), ash (2 hrs at 600 °C), crude lipid (Bligh and Dyer 1959),total nitrogen (Technicon Auto Analyser methods 369-75 A/A and 334-74 W/B) and gross energy (adiabatic bomb calorimeter-Gallenkamp).The amino acid compositions (Table 21a,b) were determined by anAmino Acid Analyser (Beckman 6300) at the University of BritishColumbia and by Triple A Laboratories, Seattle, U.S.A.(BeckmanSpinco Analyzer 120B). The mineral compositions (Table 22) weredetermined by plasma emission spectroscopy (ICAP) at Quanta TraceLaboratories Inc. (Burnaby, B.C.). TVN and TMA were determined bymicrodiffusion with K 2CO3 as described by Beatty and Gibbons (1937)and as modified by Murray and Gibson (1972) and Malle and Tao(1987) and they are expressed as mg nitrogen/100 g sample (dryweight). PV and TBA (method B for fish meal) values weredetermined as described by Woyewoda et al. (1986).Pepsin digestibility was determined by the modified Torrymethod (pepsin activity 1:10,000 at .0002% concentration) asdescribed by Lovern (1965).The pH-stat method employed in the present study measures thedegree of hydrolysis (DH) of the sample protein at 15°C for 500minutes by using digestive enzymes from the pyloric ceca of rainbowtrout with the aid of an auto-titrator system interfaced to a PCcomputer. The method has been used to determine the digestibilityof a wide range of marine protein sources (Haard, 1991).151Table 20Proximate Composition and Gross Energy ofJumper Test Protein SourcesProtein Moisture Ash (%) Protein Lipid Energy'Source Vs) (%) 1 (901 (MJ/kg)LTF-12 11.39 5.88 66.18 26.38 26.40HTF-2 8.46 6.24 66.05 27.33 26.82LTM-3 8.68 5.87 64.42 28.65 26.61HTM-4 8.96 5.70 64.55 27.48 26.78LTS-5 10.02 5.21 60.46 31.43 28.41HTS-6 8.59 5.65 63.12 29.68 26.95RDH-7 7.72 14.37 73.17 10.75 22.00Dry Matter Basis2 Refer to Table 19 for additional information.152Table 21aAmino Acid Composition of Jumper Test Protein Sources'AMINO- . LTF -1 	 HTF -2 LTM -3 HTM -4 LTS -5 'HI'S -6ACIDS-Ala2 5.39 	 5.86 5.88 6.19 5.87 5.83Arg 6.23 	 6.49 4.58 5.49 7.10 2.73Asp 8.11 	 8.46 8.48 8.90 8.49 8.43Cys3 2.26 	 1.43 1.48 2.62 1.43 2.54Glu 11.30 	 12.02 12.11 12.73 12.15 12.15Gly 6.79 	 7.39 7.34 7.76 7.32 7.32His 1.68 	 1.83 1.77 1.92 1.86 1.85Ile 3.10 	 3.38 3.36 3.59 3.41 3.36Leu 5.57 	 6.05 6.01 6.26 6.14 6.08Lys 6.84 	 7.34 7.17 7.77 7.20 7.16Met 1.78 	 2.19 1.68 2.17 2.21 2.12Phe 1.51 	 1.70 5.13 1.78 1.63 1.65Pro 3.93 	 4.22 4.29 4.41 4.25 4.24Ser 3.83 	 3.98 4.05 4.21 3.96 3.96Tau 1.05 	 1.53 1.13 1.36 1.56 1.55Thr 3.56 	 3.69 3.83 3.88 3.73 3.72Trp3 .88 	 .67 .83 .72 .62 .79Tyr 1.92 	 2.11 2.94 2.20 2.05 2.06Val 3.70 	 4.03 3.79 4.12 4.05 3.98Total 79.43 	 84.30 85.90 88.08 85.02 81.52Expressed as g/16 g of Nitrogen (% protein).2 As determined by amino acid autoanalyser(Beckman 6300) at the Animal ScienceDepartment, U.B.C. Vancouver unless otherwise stated.3 As determined by Triple A Laboratories, Seattle, WA, U.S.A.153Table 21bAmino Acid Composition of Jumper Test Protein Sources'ACIDSAMINO : HTF-2 LMM-3 HTM-4 .HTS-6Ala2 3.58 3.87 3.79 4.00 3.55 3.67Arg 4.12 4.29 2.95 3.54 4.29 1.72Asp 5.37 5.59 5.47 5.75 5.13 5.32Cys3 1.50 0.94 0.95 1.69 0.87 1.60Glu 7.48 7.94 7.81 8.22 7.35 7.67Gly 4.49 4.88 4.73 5.01 4.42 4.62His 1.11 1.21 1.14 1.24 1.13 1.17Ile 2.05 2.24 2.17 2.32 2.06 3.84Leu 3.69 4.00 3.87 4.04 3.71 3.84Lys 4.53 4.85 4.62 5.02 4.35 4.52Met 1.78 2.20 1.68 2.17 2.21 2.12Phe 1.00 1.12 3.30 1.15 0.99 1.04Pro 2.60 2.79 2.76 2.85 2.57 2.68Ser 2.53 2.63 2.61 2.72 2.39 2.50Tau 0.70 1.01 0.73 0.88 0.95 0.98Thr 2.36 2.44 2.47 2.51 2.25 2.35Trp3 0.58 0.44 0.54 0.47 0.38 0.50Tyr 1.27 1.40 1.90 1.42 1.24 1.30Val 2.45 2.66 2.44 2.66 2.45 2.51Total 53.18 56.47 55.94 57.64 52.26 53.94Expressed as g/100 g dry weight sample.2 As determined by amino acid autoanalyser (Beckman 6300) at the Animal ScienceDepartment, U.B.C. Vancouver.3 As determined by Triple A Laboratories, Seattle, WA, U.S.A.154Table 22Mineral composition of Jumper test protein sources'ELEMENT 	 LTF-1 HTF-2 LTM-3 HTM-4	 HTS-6ANALYSISCalcium 8135.0 8960.0 8770.0 8800.0 6995.0 7580.0Chromium < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0Cobalt < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0Copper 1.5 2.5 2.0 2.5 1.5 3.0Iron 68.5 71.5 66.0 70.0 61.5 66.5Magnesium 875.0 940.0 825.0 870.0 715.0 775.0Manganese .8 .8 .75 .8 .7 .55Phosphorus 7135.0 7825.0 7485.0 7500.0 6315.0 7095.0Potassium 7450.0 6750.0 7100.0 7450.0 6350.0 7750.0Selenium 8.0 8.5 6.5 7.5 7.0 8.5Sodium 7685.0 8205.0 7115.0 7605.0 6640.0 7430.0Zinc 40.0 41.5 37.0 39.0 36.0 38.0' Expressed as mg/kg dry meal as determined by Quanta Trace Laboratories Inc.Burnaby. B.C.1556.2.2 Diet Composition and FormulationThe composition of the basal diet has been presented inTable 12. The mineral and vitamin premixes were mixed and thenadded after grinding (Fitzmill-JT with #30 U.S. screen) and mixing(Marion mixer) of the other dietary components. After furthermixing of the dry ingredients some of the herring oil was added tothe mash before cold pelleting (California Pellet Mill- 3.18 mmdie). The remainder of the oil was sprayed (electric spray gun)onto the pellets (after drying and cooling in vertical dryer andscreening-#10 U.S. screen for fines) as they tumbled inside acement mixer. The pellets were then screened again and stored ina cold, dry ingredient room until used. For the test diets, 70% ofthe basal diet mash containing part of the oil and 30% of the testprotein source were mixed in a Hobart Mixer. Pelleting was thendone as above, and the balance of the oil was sprayed onto thepellets.Chromium sesquioxide (Cr2 0 3 ) (Fisher Scientific) was added ata level of 0.5 % to all test diets and the reference diet prior tocold pelleting.6.2.3 Aquarium facility and Experimental protocolRainbow trout were purchased from a local hatchery (SpringValley, Langley) and transported to the laboratory in April, 1990.They were placed into stock tanks supplied with dechlorinated citywater and fed Biodiet twice daily to satiation for 7 days. Thefish were weighed, selected (10-18 g) and distributed randomly tothe 21 digestibility tanks as described in the chinook protocol156with 50 fish/tank. The experimental conditions are described inTable 23.Juvenile chinook salmon (monosex females) were obtained froma local hatchery (Sea Spring Salmon Farms, Vancouver Island) andthey were transported to the laboratory in May, 1990. The fishwere placed into stock tanks supplied with well water. One weekprior to transport, the fish were hand vaccinated to protect themagainst vibriosis and furunculosis (Furogen-Vibriogen-AquaHealth Inc.). They were then slowly acclimated to salt waterover a 10-day period. Fish were fed 2-3 times daily to satiationwith Biodiet. They had grown to 9-13 g by June 19/90 and wereanesthetized and distributed randomly to the 21 digestibility tankswith 50 fish/tank. The experimental conditions are described inTable 23.In both of the foregoing experiments, the fish were fedchromic oxide-free basal diet for 7 days and then they were re-weighed and fed the test diets for the remainder of the experiment.All fish were fed their prescribed diet by hand twice daily (threepasses) to satiation (approximately 2-3% body weight/day). Dailyrecords of feed consumption and mortality were maintained.Feces were collected, frozen, and freeze-dried as outlined inexperiment 3 (chapter 5).Feces and feed were analyzed for proximate composition(moisture, ash, lipid and protein) and for the chromic oxideindicator concentration as determined by the wet ashingcolorimetric method of Stevenson and De Langen (1960).157Table 23. Experimental conditions during the rainbow troutand chinook salmon studies (Jumper)Experiment no.1	 2Experimental fishInitial mean wt(g)Number of fish/tankStocking density (kg/m3 )Number of groups/dietType of tankWater volume (1)Source of waterWater temperature (°C)Water supply (1/min)Salinity (ppt)Dissolved oxygen (mg/l) b	9.9-10.3Feeding frequency/day 	 2Duration of feeding (days) 	 17Photoperiode 	 naturalStart time 	 April 24/90Chinook salmon12.02504.013Fiberglass150Salt water10.0-14.86-828-317.6-8.9219naturalJuly 6/90Rainbow trout13.72504.573Fiberglass150City watera7.6-8.26-80a Dechlorinated city waterb Monitored by OXYGUARDc Photocell to Vitalite Durotest light158The apparent digestibility coefficients were measured asoutlined in experiment 3 (chapter 5).5.2.4 Fish husbandry and measurement of growthAt the end of the experiment (17 days for rainbow trout and 19days for chinook salmon) fish were anesthetized and individual fishweights were recorded. Specific growth rate was calculated asfollows:SGR = In Final weight(g)- In Initial weight(g) x 100 duration of experiment (days)The feed efficiency ratio was expressed as:FE = 	wet weight gain (q) dry feed intake (g)The protein efficiency ratio was expressed as:PER = wet weight gain (g) protein intake(g)6.2.5 Statistical analysisThe nitrogen and organic matter apparent digestibility valueswere analyzed by analysis of variance (ANOVA). A factorial designwith two processing temperatures and three levels of spoilage inthe raw material was used to assess final fish meal quality. Thehomogeneity of variance of the data was tested with Bartlett's test(Systat 1990).The effect of species (rainbow trout and chinook salmon) ondigestibility values for organic matter and protein of the 6 dietswere compared using ANOVA. When significant differences betweenspecies were detected the mean digestibility values of each testprotein source for rainbow trout and chinook salmon were testedusing a t-test (p=0.05). Comparisons between digestibility values159(organic matter and protein) for each of the test proteins wasdetermined using Tukey's test (Systat, 1990).Correlations between the in vitro digestibility and salmoniddigestibility values and level of spoilage indicator for the testprotein sources were also determined.6.3 RESULTSThe test protein sources were all easily ground (Fitzmill).Also, they mixed well with the other feedstuffs and in every casethe resulting mash was easy to cold pellet. The consistency of theprotein hydrolyzates was similar to that of fish meal but unlikethe latter, they were very homogenous and had no large scales orbones present.The proximate compositions of the various hydrolyzates arepresented in Table 20. The protein levels in the hydrolyzatesranged from 60.5 - 66.2% of dry matter. Protein level decreased asthe level of spoilage of the raw material increased. This decreasein protein level, particularly with the highly spoiled material,was accompanied by an increase in lipid concentration. The aminoacid compositions of the test protein sources did not seem to beaffected markedly by spoilage or by the processing conditionsemployed, except in the case of HTS-6 (low Arg) and LTM-3 (highPhe) (Tables 21a,b). Similarly, the mineral compositions(Table 22) were also not affected by spoilage or heating, exceptfor LTS-5 which had reduced levels of several of the elementsrelative to other test protein sources.The spoilage indicators (TVN,TMA) increased slightly as160\ \a)tOcdOzE 	 O OCOOCiCSCO0. I\‘.X X\M‘co 	 co0 0C\J0161storage time increased (Fig. 8). No trends were evident for theTBA and PV levels present in the test protein sources (Table 24).Apparent organic matter and nitrogen digestibility in both therainbow trout and the chinook salmon were not appreciably (p >.05)influenced by temperature of processing. However, the increasedlevels of spoilage in the raw material significantly depressed (p<.05) apparent digestibility of organic matter and nitrogen in bothrainbow trout and chinook salmon (Fig. 9, Table 24).Pepsin digestibility values of the test protein sources wereall very high (>98%) and were unaffected by temperature ofprocessing and freshness of raw material (Table 24). The pH-statvalues were slightly higher for the most spoiled meals as comparedto the other meals (Table 24).Fish performance (specific growth rate, feed efficiency ratio,protein efficiency ratio) was not affected by processingtemperature or spoilage of the raw material (Table 24). Both thechinook salmon and rainbow trout actively fed during bothexperiments and were in a state of positive nitrogen balance.6.4 DISCUSSION6.4.1 Test protein compositionThe protein levels in the hydrolyzates were slightly lowerthan normally found in fish meals from Norway and Denmark (>68-70%). At the same time the levels of lipid were higher (>8-11.5%),and ash lower (<14%) than the published levels for fish meals fromNorway and Denmark (NRC, 1981; 1983; Sandfeld, 1983; Smith, 1990),although the proximate compositions were comparable to those of1620_'77al!  MM.! NM!  111111!0 0COoo00:E()0a.)C)111111111111111111111111111113iii^ir iii-NG00:EO•0:1• %ct• co13	 44 In	 IICD 	 D) •	 0•--1 	 al 0	 -.44,4O 0 •ta04 	 .73.4• 4-4 in >1 •••••••ICJ 	 cd • 0 	 ID• 0 ••0O U 0 a)	1:1 II 	 44E 14 C 0• U C• 0 n:I 	 -. .44-.4 	 4.) 	 4-$ 441:1 	 -.4 •••• CO 4-4 00) 	 •:1'E $4 Is) 4.3ro •-• 	 4.1••	 .0 • In a) 0.0 	 0 r-I • la>024 04a) 	 - ■-10 	 0• EI 	•144 	 0 	 •-•c0 E131a) 	 • •-• 	 •-•Ut 	 CO 0 	 (0 00 	 E' 	 4-1a) 0c/3 	 'CS $4 a) '0O (a w 00 	 a) --I 4.)4-1ca 	 a) a) 4.4 a)cri 	 04 43- oX 	 0 $4• >1• ±j• 	.-1 411g II114 	 ID 0 .L.1 4  ----13 	 0 	 0cn'0 0X 0 	 -41E04 0 al$4 	 44 °o 344.10)	c1411) a)34 	 H>4	4.) a	 c.)4.) 	 an:I (I)- •-1•	04 E 4-1	4-1 	C.)COO	 •• C.) 0 .5'CD 	CID -.4 0O -.4 al a)- a)E• s4".-04 4.1al )4 	 „..1▪ 0.0 	 >4 	 4-1	04 	 4-• 1-.4 	 4) 4.) -.4• 0CI) C 	 $4E 0 	 041:1 	 $4 IDO $4 4.) $4ro 	 E4 	 f=43•II 	 X••-•04  01 CT)04(1) 	 4.)3	 •0o o0al 00 Ez 	 a.)•■•1 0 ..4E CT .0 C.)X 0 	 C.)$4 n:I4.) 	 1/440 4C 44 4.1	 1	 1:$,-I 	a)o a) 3 4..)4.) id 44**-1 	 (1)U) 4-i oLI 0 $4 .0 	 $441$ $4 4.1 .044a)0 0>).1.0CD a) 	 131ID • 04•1 	 'CICTC U•-•0 of 	 0 	 1/44 C3-iE-■ 71 to 	 .4-4,1 -ftl4-1	(1.1	 II C C U 01al 4-1 Z 0 0 al4-) a) 	 04 4-10 E Z CO 3•-■	 rg en •t163v;)1toE4WDI s in el1-1 1.0 N	 •• • 	 • TrCO 10 r.410d' r-I H 1•1elc..)CAONCl•1"---,-)crelCOal2 	 . 	 2 ..M .4,	0 COv-I 01 	 01 01• • 	 • 	 •0 N	 CI NCY1 01 	 01 01I-I 01 CI 	 111 •4, 0CI N sz? 	 10 r) 0tO N 1/40	 CA HV)• • 	 • 	 • 	 • 	 •	1- 	 1-1U1iCaE-1641/40 1/40 CO 0v o N •• • 	 • 00tc) tn o co	Tr 	 r•1 	 ,--1 	 ..-10111coNN•Nr-1U1Ncoalco o 	 el kou .0 	 0 .010 In	 CT 01• • 	 • 	 •r--- up 	 0 1.0CO 01 	 01 CA	CI N CO 	 01 CO cr)	01  CI 	 el N 01	1/40 v-1 tO	 01 	 1 1.01• • 	 • 	 • 	 • 	 ••--) 	 •--1%IIIZE4M01 Cf) 0 00 0 CO 	 •• • 	 • 	 kipU1 Ul •-•I ct,Tr I-I 1-1 r-IelM,.„''d'r)tr.)r-1Chmis01	CA 111 	 •r-I N	CO 1/40 	 U) C)• • 	 •	 •	N CO	 01 t•-•	0101 	 0101H CO tf)	 03 0 COc-- o cn 	 U") .1, LC)V) v-1 U)C/1 N lf)• •	 •	 • 	 • 	 •	1- 	•-1inIZE-41-101 N 1•1 •c1'0 c0 m 	 •• • 	 • 	 kr)rI 4.1 N InTr 1-4 I-I •-1r-,-„.,‘....`"''''''C)0•1/40r-Ir)00101a 	 a 	 a 	 a1/41:) in	 N NN In 	 • 	 •• • 	 ,-1 	 ,-.4U1 CO	 0 001 01 	 •-1 1-101 N C)	 H 0N N r-1 	 10U) 1-1U") N CO 	 01 11 10• • 	 • 	 • 	 •• I- 	 r4NI44ElWNr-ION1/413 N 	 •zr 	 •• • 	 • alo cn in en'd' 	 1-1 	 1-.1 	 1-1r)C1_`" '01NC3v-IIr)01Cocs	a 	 a 	 II	0 • I 	 al .1.	 1 0 	 • in• • 	 0,1 	 •	el N	 001	ol 01 	 r-4 01,--1 Lc) Tr	 al vr o0 H N	 0 •cl. •cr10 N 1.11 	 01 N /-I3• • 	 •	 •	 • 	 ••-I 	 1-11-1IIi/EtAU1 01 I-I 01ON N 1/40	 •• • 	 • 	 1/40oN o Nr) ri r-t t-i0in,..,-...0Md'r-11/40Nco011 	 .0.0 	 ••	 ••	 .01-1 	 1/4(;) 	 ko 	 cor-1 0	 CON• • 	 • 	 •N00 	 10N01 01 	 Ch 01	I-I 0  / 0	 01 r) 1-1	tr) cc, H	 o t-t N	 n C)	 al <-4 ko• • 	 • 	 • 	 • 	 •	1-1 	 1-1-E.)03H0,..4 	 ...-■ .••••■ 	•04 tr, 01 	- •0 	 tpE 0 0 	 gicc 0 0 - --...,C.) 	 r-I r•I	 triH ---.. ---.. ,.	 C110 	 .....,Q)Z Z Z 	 I-1H 	 erl 00101 w 0ralgEZ0 '-4A Z 4 	 4H > X > ciaCUCUCA•...C.,)0---•040a)EtIT0•rI>1I•40OP..ZCI--1.3el4.1inIWCli•-••0*•ri •-•GIa>10 0r I 4 • r Igi0 ••••10 Ao0I-1 I-4-1-1 00 cl)>I 	 '''.1EA 	 W 	 WH 	 4-) 	 .0A Z 4-) 	 4-1H 0 al 	 E-1 ICSAl X 0	 0H 1-1 CO CE ic4 0 a) P4 U Wta En ' ,'"I tri E-1 ,--1 01ral 	 C 0	 = 0CDXITI3-103$.1HOZTPOCP4-)ISI 0 	 I-1 • ,-1 cla 	 P • ,-1ZOZZOZZ H 	 HCO Z1-4 C_) 	 grx■	4 >1 	 4 >1	 -3 0	 4-1 0	Z 	 Zo14 	 w 	 o a)	Ci Z 3  -H 	 $4 •-1	0 t 	 Ei tri c.)	... 	 -i-i	5 0 -1 	 0 0 4-1	KIC . ,4-1 	 04 - I-1 (4-1	0 co 4-4 4	 E-14-141rig, 	 -r-1 	 -.4- x 0'0 3 0 '0WOW WZOWW04	12	 CLI cll 	 1.74 CO 	 fa• w	 r4Z co PL. o4 Z cn 44 o4Z H 	 HU) ZH Urz■164some other fish protein concentrates (Hale, 1972; Stone and Hardy,1986). Protein levels decreased as the extent of spoilageincreased in the test protein sources. Lassen et a/. (1949, 1951)obtained similar results in spoiled condensed fish solubles andfish meal and attributed the decreased levels of protein tobacterial action and to a change in the nitrogen distribution(higher levels of volatile nitrogen, ammonical nitrogen and othernon-protein nitrogen). These volatile components would have beenlost, thereby resulting in a lower level of nitrogen and thus,percent crude protein (%Nx6.25). Stone and Hardy (1986) obtainedsimilar results, with ammonia levels increasing in various fishsilages and length of polypeptide fractions decreasing during a 42-day storage study.Fish, like other animals, do not have a true proteinrequirement (Wilson, 1989) but require a well balanced mixture ofessential and nonessential amino acids (Higgs et al., 1988)(Table 1). The amino acid composition of the test protein sourceswas similar to that of white fish and South American type fishmeals (FAO, 1986; Asgard, 1988). Most of the protein sources metthe essential amino acid requirements of rainbow trout, with theexception of arginine (low in HTS-6) and tryptophan (low in allsources). Similarly, most of the essential amino acid requirementsfor chinook were met, with the exception of arginine (low in threesources), cystine plus methionine (low in three sources) andphenylalanine plus tyrosine (low in five sources). Tryptophanlevels may have been reduced as a result of the hydrolysis process,165since it has been shown that short term hydrolysis (<17 hrs)results in low tryptophan levels because the release of tryptophaninto the soluble phase is slower than that of other amino acids(Hale, 1972). The amino acid composition of the hydrolysates didnot seem to be affected by the extent of raw material spoilage orby the processing conditions employed in meal production. Thisfinding is contrary to those of Lassen et al. (1949) who foundconsiderable destruction of amino acids by bacterial putrefactionof condensed fish solubles, with the destruction of valine beingparticularly high. Sandfeld (1983) likewise noted decreasedlevels of amino acids in fish meals made from previously spoiledraw material. Also, he found that high drying temperature duringmeal production greatly decreased the levels of thermo-labile aminoacids (arginine, cystine, lysine, methionine, serine andthreonine). Such destruction, however, was not apparent in thepresent test protein hydrolyzates probably because there was notextensive spoilage of the raw material and relatively lowprocessing temperatures were used in the hydrolysis process.The mineral composition of the test protein sources did notseem to vary with processing temperature or freshness of the rawmaterial, except for LTS-5 which had lower levels of most mineralsthan the other protein sources. The mineral requirements forrainbow trout, Atlantic salmon, and chinook salmon are presented inTable 2. It is difficult to define requirements of individualminerals because of the complex interactions which occur betweenminerals, enzymes, vitamins, lipids and proteins (Tacon, 1985;166Richardson et al., 1986; Lall, 1989). Generally, high dietarylevels of phosphorous and calcium accompanied by low levels of zinccan lead to high incidence of cataracts in both rainbow trout andchinook salmon if mineral binding agents are simultaneously presentin the diet. If dietary zinc levels are high, the incidence ofcataracts is greatly reduced (Richardson et al., 1985; 1986). Thelevels of minerals in the test protein sources did not exceed thereported toxicity levels for salmonids (Tacon, 1985).6.3.2 Spoilage IndicatorsThe levels of TVN and TMA increased only slightly in the mostspoiled protein sources (LTS-5, HTS-6)(Table 24) and therefore,these parameters were only of limited utility in predicting thequality of fish hydrolyzates that were only moderately spoiled,even though TVN level in raw material has been used to determinethe quality of fish meal. For Norwegian LT special meals, it hasbeen stipulated that the level of TVN in the raw material shouldnot exceed 90 and 50 mg /100 g fish for Norsemink and Norse-LT 94,respectively. For Danish fish meals the levels of TVN should notexceed 301 and 134 mg N/100 g meal for standard and special fishmeals respectively (Pike et al., 1990; Smith, 1990). Problemsarise in comparing TVN and TMA values, as the methodology for theTVN and TMA analyses can greatly affect the measured levels (Bottaet al. 1984; Clancy et al., 1990). Care should therefore be takenwhen expressing TVN and TMA values to state the methodology used.Studies have been undertaken to standardize the tests for TVN andTMA in fish products (Antonacopoulos and Vyncke, 1989).167The TVN and TMA levels in the test protein sources increasedslightly in the final product when the raw material had been heldfor 8 days, but heating had little effect on these levels. Thiscontrasts with the findings of Sandfeld (1983) who discovered thatTVN levels decreased in the fish meals processed at highertemperatures (>100 °C). He attributed this to the loss of volatileamines in the press liquor at higher temperatures. Pike et al.(1990) reported similar findings in which the levels of biologicalamines in fish meals increased with spoilage but decreased withheating (>140 °C). Perhaps the higher temperature (93°C) used inthe hydrolysate process and the nature of the product resultingfrom enzymatic hydrolysis might have lowered the TVN and TMA levelsduring processing of the raw material. The present results supportthe work of Halland and Njaa (1989) who concluded that TVN was oflittle value for assessing the quality of fish silage which is alsoproduced by an enzymatic process.TVN and TMA levels for the Jumper meals correlated poorly andnegatively with organic matter and nitrogen digestibility valuesfor both species. The highest correlation (r= -0.649) was betweenchinook organic matter digestibility and TVN levels. Although, thelevels of TVN in the meals correlated highly (r= 0.883) with TMAlevels (Table 25).Oxidation is a free-radical process proceeding throughinitiation, propagation and termination steps and yields aldeyhdes,acids, epoxides, diglycerides, monoglycerides and polymers.However, since end products are continually being lost to further168Table 25Correlations between various digestibility values and spoilageindicators of Jumper test protein sources for pH-stat, pepsinsolubility, organic matter and protein digestibility with rainbowtrout and chinook salmon.FISH VS FISHRT-OM vs CHIN-OMRT-NI vs CHIN-NIR R2- 	 .653 .427- 	 .761 .580- 	 .653 .427- 	 .516 .266- .507 .257.916 .839.795 .633.526 .277.804 .647.951 .904.594 .353-	 .627 .393- 	 .649 .422- 	 .592 .351- 	 .276 .076- 	 .349 .122-	 .364 .132- 	 .206 .042-	 .170 .029.883 .780IN VITRO VS FISHpepsin vs OM-RTvs OM-CHINvs NI-RTvs NI-CHININDICATOR VS FISHTVN vs OM-RTTVN vs OM-CHINTVN vs NI-RTTVN vs NI-CHINTMA vs OM-RTTMA vs OM-CHINTMA vs NI-RTTMA vs NI-CHININDICATOR VS INDICATORTVN vs TMAComparisonIN VITRO VS IN VITROpH-stat vs pepsinIN VITRO VS FISHpH-stat vs 0M-RT Ivs OM-CHINvs NI-RTvs NI-CHIN1 RT= Rainbow trout, CHIN= Chinook, OM= Organic matter, NI=Nitrogen, TVN= Totalvolatile nitrogen, TMA= Trimethylamine169reaction, measurement of the extent of oxidation is difficult. Thethree most popular laboratory methods applied to the measurement ofoxidation in fishery products, namely TBA, PV, and carbonal value(COV), are of a reduced value because of the transitory nature ofproducts measured, which increase to a maximum and then decline.Therefore it has been recommended by Woyewoda et al. (1986) that atleast two tests be used to measure the extent of oxidation.The values for PV in the test protein sources did not presenta clear picture (Table 24). The TBA values were lower for hightemperature meals than those for the low temperature meals (forpaired meals-same level of spoilage in each batch). 	 Therecommended value for PV and TBA in fish oils is < 10 meq /kg and< 70 mg malonaldehyde /kg (NRC, 1981). Thus, the PV values seemrather high (10.38-15.4 meq /kg) for the test protein sources.Ketola et al. (1989) found that growth rates of Atlantic and cohosalmon fry were impaired when they were fed diets containing highlyrancid oil. In the present experiment it was not known whether theTBA and PV values would have been increasing or decreasing at thetime of sampling, because of the nature of the product and storageconditions (stored at 15°C for several months before dietpreparation and analysis). El-Lakany and March (1974) found thatTBA and PV values initially increased and then decreased over timein freeze-dried herring stored at various temperatures.6.4.3 Fish performance and digestibilityThe specific growth rates of the rainbow trout and chinooksalmon were unaffected by spoilage and processing conditions of the170test protein sources. The specific growth rates for rainbow troutwere higher than noted for chinook salmon even though the chinooksalmon were in warmer water during the experiment. This may havebeen due to many factors such as species differences, dissimilarcultural conditions and possibly the level of domestication oftrout relative to chinook salmon. Further, the chinook salmon mayhave been more stressed in the digestibility tanks than the rainbowtrout (Hajen, 1990). Feed efficiency ratio and PER were alsounaffected by spoilage and processing of the test protein sourcesin both chinook salmon and rainbow trout. The rainbow trout,however, had higher specific growth rates than chinook salmon. Thegrowth performance of each species was only of limited use in thissituation for screening feedstuffs for potential nutritive valuebecause of the short duration of the experiments and thepossibility of fish stress in the digestibility tanks due to dailycleaning of filters and flushing of the tanks. Also, the dietsused in experiments such as these cannot all be balanced to exactlymeet estimated nutrient requirements. For more accurate growthdata, longer duration feeding trials and multiple levels ofinclusion of test protein sources in balanced diets arerecommended.The organic matter and protein digestibility values for thetest protein sources in both the rainbow trout and chinook salmonwere not significantly (p > 0.05) influenced by the higher dryingtemperature. By contrast, other studies have shown decreaseddigestibility in salmonids, mink, swine and poultry fed diets171containing high temperature processed (>100 -140 °C) fish meals(Tarr and Biely, 1973; Sandfeld, 1983; Opstvedt et al., 1984; Pikeet al., 1990). The processing temperatures used in this study werewithin the acceptable range for high fish meal quality (Pike etal., 1990).Apparent digestibility values for protein and organic matterin the test protein sources for both chinook salmon and rainbowtrout were influenced by spoilage of the raw material. As spoilageof the raw material increased, digestibility decreased forsalmonids fed the test protein sources (Fig. 9). Organic matterdigestibility values for the test protein sources obtained withrainbow trout correlated highly (r= 0.951) with those valuesobtained with chinook salmon but the trend was less apparent fornitrogen digestibility values (r= 0.594) (Table 25).Organic matter digestibility was higher for rainbow trout,particularly for HTF-2 which was significantly higher (p=0.05) thanfor chinook salmon and it was influenced more than nitrogendigestibility by spoilage of the test protein sources. Thissuggests that lipid digestibility may have been affected more thanprotein digestibility and that chinook salmon are even moreadversely affected by spoilage of raw ocean perch than rainbowtrout. Nitrogen digestibility for LTM-3 obtained by rainbow troutwas significantly different (p=0.05) than for chinook salmon.Austreng (1988) found that Atlantic salmon were affected more thanrainbow trout when fed acidified spoiled silages. Stone and Hardy(1986) also noted depressed digestibility of silages stored for172long periods of time when fed to rainbow trout and attributed thisto a decrease in lipid as well as protein quality.The protein digestibility of the reference diet was higher inthe chinook salmon than in the rainbow trout. Even though the samereference diet was fed to both salmonids the apparent nitrogendigestibility value was 10% lower for rainbow trout. This may havebeen the result of species differences and/or the influence ofsalinity. Lall (1988 as cited in Higgs et al., 1992) found thatAtlantic salmon in sea water had higher protein digestibilitycoefficients for several feedstuffs than Atlantic salmon in freshwater. In contrast to this, Usher et al. (1990) found a decrease(96.3 to 87.4) in apparent protein digestibility of the overalldiet following transfer of Atlantic salmon into sea water. Theyexplained the reduced protein digestibility as a consequence ofincreased evacuation rate in the gut of fish at a higher salinity.It is unknown whether salinity has any effect on the gastricevacuation rate in chinook salmon. A possible explanation for theincreased protein digestibility by chinook salmon in sea water isthe activation of pepsin I by NaCl. Sanchez-Chiang et al. (1987)found that chum salmon have two pepsins (I, II) in their stomachsbut both pepsins are only activated in fish in salt water. It isnot known if a similar isoenzyme system exists in chinook salmon.Pepsin digestibility was very high (>98%) for all test proteinsources regardless of processing temperature or degree of spoilageof raw material. Pike et al. (1990) found that markedly higherprocessing temperatures (140°C) for fish meal decreased pepsin173digestibility, while spoilage of raw material increased pepsindigestibility. The highest correlation (r= 0.916) was found betweenpepsin solubility and organic matter digestibility with rainbowtrout for the Jumper meals (Table 25).The pH-stat values were highest for the most spoiled meals butincreased processing temperature decreased the degree of hydrolysis(DH) most notably in HTF-2 and HTM-4 which were produced from freshand medium spoiled material. The test correlated best with organicmatter digestibility obtained with rainbow trout but thecorrelations were low and negative (Table 25).The organic matter and protein digestibilities of even themost highly spoiled protein sources fed to salmonids in this studywere greater than the digestibility values reported for specialNorwegian LT meals. Pepsin digestibility for special meals shouldexceed 94%, while protein digestibilities should range from 80.5 to85.8% for salmonids (Pike et al., 1990; Smith, 1990). The apparentprotein digestibility coefficients obtained for the test proteinsources fed to rainbow trout in the present study were even higherthan the apparent protein digestibility coefficient (91.5) foundfor fish protein concentrate in rainbow trout by Lall andBishop (1977). The apparent nitrogen digestibility coefficientobtained with sea bass (99.6%) fed a protein hydrolyzate(Spyridakis et al. 1988a) and with rainbow trout (95.0%) fed a fishprotein concentrate (Cho and Kaushik, 1990) however, werecomparable to the nitrogen digestibility coefficients obtained inthe present study with rainbow trout and chinook salmon.1746.5 CONCLUSIONSAll protein hydrolyzates, except when prepared from materialheld for 8 days, were very digestible for both rainbow trout infresh water and chinook salmon in sea water. The holding of theraw material for 8 days at 2-5°C prior to processing had an adverseeffect on the digestibility of the test protein sources. Storagefor 4 days at the same temperature had little effect. Elevation ofthe processing temperature from 85°C to 93°C did not adverselyaffect digestibility. The high digestibility of these proteinhydrolyzates makes them attractive for inclusion in diets designedto minimize water pollution. The processing of coarse groundfishthat may have been at one time discarded is also of ecologicalbenefit.175Chapter 77.0 SUMMARY AND CONCLUSIONSThis study was undertaken to assess the nutritional value offish meals (Pacific herring) and protein hydrolyzates (ocean perch)processed from fresh, moderatly fresh and advanced spoiled rawmaterials at two processing temperatures for chinook salmon in saltwater and for rainbow trout in fresh water.This thesis, consisting of four experiments, was designed toinvestigate spoilage and thermal processing parameters of the rawmaterial on the quality of the final product. Experiment I(preliminary study) investigated the rate of spoilage of wholePacific herring held at 2-5°C for up to 15 days as assessed by thelevels of total volatile nitrogen (TVN) and trimethylamine (TMA)using several methods. Experiment II extended the results ofexperiment I and included other chemical indicators of spoilagesuch as putrescine, cadaverine, and histamine and also, evaluatedhow the levels of these indicators would be influenced by thermalprocessing. Experiments III and IV were designed to query theeffects of three levels of raw material freshness and twoprocessing temperatures employed during herring or ocean perch mealproduction on the quality of the final product as assessed bydigestibility values in both chinook salmon and rainbow trout. Invitro digestibility values were also determined using the pepsintest and the pH-stat multi-enzyme test.In experiment I, the levels of TVN and TMA increased rapidlyin herring after 8 days of storage at 2-5°C. The TVN values were176significantly different depending on the method of analysis and theTMA values also varied depending on the method used. Directdistillation with MgO yielded significantly higher TVN results thanmicrodiffusion with K 2CO3 or KOH and steam distillation with NaOH.The TMA values were highest for steam distillation with NaOH andtherefore, for the routine determination of TVN and TMAmicrodiffusion with K 2CO3 is recommended.Similarly, in experiment II, the levels of putrescine andcadaverine in herring rose sharply after 8 days of storage at 2-5°C, with large amounts of cadaverine produced by day 12. Histaminelevels remained low (< 5 ppm) during storage, which may have beencaused by pre-freezing of the herring before chill storage. Thelevels of putrescine, cadaverine, histamine, TMA and TVN were lowerin the herring press cake meals probably due to losses in the pressliquor and thermal degradation. High processing temperatures hadlittle effect on indicator levels but they did result in increasedyields of press liquor. Spoilage of the raw material also increasedthe yields of press liquor and decreased the yields of fish meal.The levels of TVN, TMA, putrescine, cadaverine and histamine appearto be useful in determining spoilage of the raw material but onlythe less volatile amines (putrescine, cadaverine and histamine) mayprove useful in predicting the quality of fish meal.In experiment III, only advanced spoilage of the herring priorto processing resulted in decreased organic matter digestibilityand digestible energy values in chinook salmon and decreaseddigestibility values for all parameters in rainbow trout in177comparison to meals produced from fresh and moderately spoiled rawherring. Elevation of the processing temperature from 75 to 100°Chad no adverse effect on digestibility values. Species differencesfor digestibility values were detected for freeze-dried herring andmeals produced from highly spoiled raw material. Lipid oxidationwas thought to decrease the lipid and organic matter digestibilityvalues, especially in chinook salmon recently transferred to seawater. Pepsin digestibility values did decrease for the hightemperature and freeze-dried herring meals but spoilage of rawmaterial resulted in elevated values for pepsin solubility. The pH-stat (enzymes extracted from rainbow trout pyloric cecae) valueswere also elevated in meals produced from spoiled raw material.Cadaverine levels in the herring meals correlated well with organicmatter digestibility values in rainbow trout. The simultaneousdetermination of putrescine cadaverine and histamine may allow forprediction of fish meal quality. This is because the levels ofpolyamines were negatively correlated with the digestibility of thefish meals in the salmonids.In experiment IV, digestibility patterns were similar to thosefound in the herring study, as spoilage of the ocean perch (storagefor 8 days) prior to thermal processing depressed the organicmatter digestibility values in rainbow trout and chinook salmonrelative to hydrolyzates produced from fresh and moderately spoiledraw material. Elevation of the processing temperature from 85°C to93°C did not adversely affect digestibility values in eitherspecies. Protein digestibility values were not influenced by178spoilage of the raw material as much as organic matterdigestibility values were. This suggests that lipid digestibilitymay have been adversely affected by spoilage. Species differenceswere also noted for organic matter digestibility values. Pepsindigestibility values were very high (>98%) for all proteinhydrolyzates regardless of the level of spoilage of the rawmaterial and the processing temperatures employed. The pH-statvalues increased slightly in the hydrolyzates produced from veryspoiled raw material. The TVN and TMA levels increased slightly inthe hydrolyzates produced from the most spoiled raw material, whilethiobarbituric acid values and peroxide values varied erraticallyin the protein hydrolyzates.It is concluded that advanced spoilage of raw material priorto processing is undesirable due to losses in fish meal yield andof quality as assessed by digestibility trials with rainbow troutin fresh water and chinook salmon in salt water.179REFERENCESAbabouch, L., Afilal, M.E., Benadbeljelil, H. and Busta, F.F.,1991. Quantitative changes in bacteria, amino acids andbiogenic amines in sardines (Sardina pilchardus) stored atambient temperatures (25-28°C) and in ice. Int. J. Food Sci.Tech. 26: 297-306.Aitken, A. and Connell, J.J. 1979. Fish. In: Effects of Heating onFoodstuffs. (Ed.) R.J. Priestly, Applied Science Publishers,London, pp. 219-254.Aksnes, A. 1988. Location of enzymes responsible for autolysis inbulk stored capelin (Mallotus villosus). J. Sci. Food Agric.44: 263-271.Aksnes, A. 1989. Effect of proteinase inhibitors from potato inquality of stored herring. J. Sci. Food Agric. 49: 225-239.Aksnes, A. and Brekken, B. 1988. Tissue degradation, amino acidliberation and bacterial decomposition of bulk storedcapelin. J. Sci. Food Agric. 45: 53-60.Analytical Methods Committee. 1979. Recommended general methodsfor the examination of fish and fish products. Analyst 104:434-450.Anderson, J., Capper, B.S. and Bromage, N.R., 1991. Measurement andprediction of digestible energy values in feedstuffs for theherbivorous fish tilapia (Oreochromis niloticus Linn.). Brit.J. Nutr., 66: 37-48.Antonacopoulos, N. 1985. Buchi- distillation unit 315 modifiedfor TVB-N determination. Buchi (Brinkmann).Antonacopoulos, N. and Vynke, W. 1989. Determination ofvolatile basic nitrogen in fish: a third collaborative studyof the West European Fish Technologists Association (WEFTA).Z. Lebensm Unters Forsch. 189: 309-316.Arnold, S.H. and Brown, W.D., 1978. Histamine (?) toxicity fromfish products. Adv. Food Res. 24: 114-154.Asgard, T. 1988. Nutritional value of animal protein sources forsalmonids. Aquaculture International Congress and Exposition,Sept. 4-7, 1988, Vancouver, B.C. pp.411-418.Austreng, E. 1978. Digestibility determination in fish usingchromic oxide marking and analysis of contents from differentsegments of the gastrointestinal tract. Aquaculture 13: 265-272.180Austreng, E. 	 1988. Acidified feed silage for fish: Practicalexperiences. Can. Tran. Fish. Aquat. Sci. No. 5347. 11 pp.Austreng, E. and Gjefsen, T., 1981. Fish oils with differentcontents of free fatty acids in diets for rainbow troutfingerlings and salmon parr. Aquaculture 25: 173-183.Austreng, E., Skrede, A. and Eldegard, A., 1979. Effect of dietaryfat source on the digestibility of fat and fatty acids inrainbow trout and mink. Acta Agric. Scand. 29: 119-126.Austreng, E., Skrede, A. and Eldegard, A., 1980. Digestibility offat and fatty acids in rainbow trout and mink. Aquaculture 19:93-95.Bachrach, U., 1973. Function of Naturally Occurring Polyamines.Academic Press, New York. p. 211.Bagnasco, S. 1985. Quantitative determination of methylaminesusing microelectrodes. Anal. Bioch. 149: 572-574.Baranowski, J. and Pan, B.S., 1985. 	 Effect of freezing andsubsequent thaw abuse. In: Histamine in Marine Products:Production by Bacteria, Measurement and Prediction ofFormation. (Eds) B.S. Pan and D. James. FAO Fish Tech Paper#252: 36.Baranowski, J.D., Frank, H.A., Brust, P.A., Chongsiriwatana, A. andPremaratne, R.J., 1990. Decomposition and histamine content ofmahi mahi (Coryphaena hippurus). J. Food Prot. 53: 217-222.Barlow, S.M. and Windsor, M.L., 1984. Fishery by-products. Int.Assoc. Fish Meal Manuf. No. 19, p. 21.Beatty, S.A. and Gibbons, N.E. 1937. The measurement of spoilagein fish. J. Biol. Bd. Can. 3: 77-91.Behling, A.R. and Taylor, S.L., 1982. 	 Bacteria histamineproduction as a function of temperature and time ofincubation. J. Food Sci. 47: 1311-1314, 1317.Bell, M.V., Henderson, R.J. and Sargent, J.R., 1986. The role ofpolyunstaturated fatty acids in fish. Comp. Biochem. Physiol.83B: 711-719.Bender, A.E. 1972. Processing damage to protein food: A review. J.Fd. Technol. 7: 239-250.Benoit, G.J. and Norris, E.R., 1942. Effect of formaldehyde on thevolatization of ammonia, mono-, di- and trimethylamines.Ind. Eng. Chem. 14(10): 823-825.181Billon, J., 011ieuz, N. and Tao, S.H. 1979. Etude d'une nouvellemethode de dosage de l'azote basique voltail total (ABVT)pour l'evaluation qualitative des produits de la peche. Rev.Tec. Vet. Ali. 149: 13-17.Bissett, H.M. and Tarr, H.L.A., 1954. 	 The nutritive value ofherring meals. 2. Effect of heat on availability of essentialamino acids. Poultry Sci. 33: 250-254.Bjarnason, J. and Carpenter, K.J., 1970. Mechanisms of heat damagein proteins:chemical changes in pure proteins. Brit. J. Nutr.24: 313-329.Bjeldanes, L.F., Schutz, D.E. and Moris, M.M., 1978. 	 On theaetiology of scombroid poisoning: cadaverine potentiation ofhistamine toxicity in the guinea pig. Food Cosmet. Toxicol.16: 157-159.Bligh, E.G. and Dyer, W.A. 1959. A rapid method of total lipidextraction and purification. Can. J. Biochem. Physiol.37:911-917.Bligh, E.G., Shaw, S.J. and Woyewoda, A.D., 1988. Effects of dryingand smoking on lipids of fish. In: Fish Smoking and Drying(Ed.) J.R. Burt, Elsevier, New York. p. 41-52.Blonz, E.R. and Olcott, H.S., 1978a. Effects of orally ingestedhistamine and/or commercially canned spoiled skipjack tuna onpigs, cats, dogs and rabbits. Comp. Biochem. Physiol. 61C:161-163.Blonz, E.R. and Olcott, H.S., 1978b. Canned spoiled tuna on growthin young Japanese quail. J. Food Sci. 43: 1390-1391.Bodwell, C.E., 1977. Problems associated with the development andapplication of rapid methods of assessing protein quality.Nutr. Rep. Int. 16: 163-178.Borgstom, G., 1955. Microbiological problems of frozen foodproducts. Adv. Food Research 6: 163-228.Botta, J.R., Lauder, J.T. and Jewer, M.A., 1984.	 Effect ofmethodology on total volatile basic nitrogen (TVBN)determination as an index of quality of fresh Atlantic cod(Gadus morhua). J. Food Sci. 49: 734-736, 750.Bowen, S.H., 1978. Chromic acid in assimilation studies. A caution.Trans. Am. Fish. Soc. 107: 755-756.Brett, J.R. and Groves, T.D.D., 1979. Physiological energetics.In: Fish Physiology, Vol. 3. Bioenergetics and Growth. (Eds)W.S. Hoar, D.J. Randall and J.R. Brett. pp. 279-352. Academic182Press, New York.Buddington, R.K. and Diamond, J.M. 1987. Pyloric ceca of fish: anew absorptive organ. Am. J. Physiol. 252: G65-G76.Bullard, A.F. and Collins, J. 1979. An improved method to araIymtrimethylamine in fish and the interference of ammonia anddimethylamine. Fish. Bull. 78(2): 465-473.Calcott, P.H., 1978. Freezing and Thawing Microbes. MeadowfieldPress Ltd., Bushey, England.Carpenter, K.J., Morgan, C.B., Lea, C.H. and Parr, L.J., 1962.Chemical and nutritional changes in stored herring meal. 3.Effect of heating at controlled moisture contents on thebinding of amino acids in freez-dried press cake and relatedmodel systems. Brit. J. Nutr. 16: 451-465.Carpenter, K.J., Lea, C.H. and Parr, L.J., 1963. Chemical andnutritional changes in stored herring meal. 4. Nutritionalsignificance of oxidation of the oil. Brit. J. Nutr. 17: 151-169.Castell, C.H. 1970. Current status of the TMA test as a measureof spoilage in fish. Fish. Res. Bd. Can. New Ser. Cir. No.38: 16-20.Castell, C.H., Neal, W. and Smith, B. 1970. Formation ofdimethylamine in stored frozen sea fish. J. Fish. Rs. Bd.Can. 27: 1685-1690.Castell, C.H., Smith, B. and Neal, W. 1971. Production ofdimethylamine in muscle of several species of gadoid fishduring storage, especially in relation to presence of darkmuscle. J. Fish. Res. Bd. Can. 28: 1-5.Castell, C.H., Neal, W.E. and Dale, J. 1973. Comparison ofchanges in trimethylamine, dimethylamine and extractableprotein in iced and frozen gadoid fillets. J. Fish. Res. Bd.Can. 30: 1246-1248.Castell, C.H., Smith,B. and Dyer, W.J. 1974. Simultaneousmeasurement of trimethylamine and dimethylamine in fish, andtheir use for estimating quality of frozen-stored gadoidfillets. J. Fish. Res. Bd. Can. 31: 383-389.Chang, G.W., Chang, W.L. and Lew, K.B.K. 1976. Trimethylamine-specific electrode for fish quality control. J. Food Sci.41: 723-724.Charalambous, G. 1986. Handbook of Food and Beverage Stability.Academic Press Inc. Orlando, Florida. 351 pp.183Chen, S.C. and Fisher, H., 1983. Gastric histamine metabolism andacid secretion in rats as influenced by diet and nutrientcontent. J. Nutr. 113: 379-387.Chettleburgh, P., Travis, D. and Pirquet, K. T. 1991. 1990-1991Statistical review. northern Aquaculture 7(3): 20-24, 40.Cho, C.Y. 1990. Fish nutrition, feeds and feeding: with specialemphasis on salmonid aquaculture. Food Rev. Int. 6: 333-357.Cho, C.Y. 1992. Feeding systems for rainbow trout and othersalmonids with reference to current estimates of energy andprotein requirements. Aquaculture 100: 107-123.Cho, C.Y. and Kaushik, S.J. 1985. Effects of protein intake onmetabolism and net energy values of fish diets. In: Nutritionand Feeding in Fish. (Eds) C.B. Cowey, A.M. Mackie and J.G.Bell. pp. 95-118. Academic Press, London.Cho, C.Y. and Kaushik, S.J. 1990. Nutritional energetics of fish:energy and protein utilization in rainbow trout (Salmogairdneri). World Rev. Nutr. Diet. Karger, Basel 61: 132-172.Cho, C.Y. and Slinger, S.J. 1979. Apparent digestibilitymeasurement in feedstuffs for rainbow trout. In: Proc. WorldSymp. on Finfish Nutr. and Fishfeed Technology Vol. II. (Eds)J.E. Halver and K. Tiews. Heenemann, Berlin. pp. 239-247.Cho, C.Y., Slinger, S.J. and Bayley, H.S. 1982. Bioenergetics ofsalmonid fishes: Energy intake, expenditure, andproductivity. Comp. Biochem. Physiol. 73B:25-41.Cho, C.Y., Cowey, C.B. and Watanabe, T., 1985. Finfish Nutritionin Asia: Methodological approaches to research anddevelopment. IDRC, Ottawa, Ont. 154 pp.Choubert, G., de la Noue, J. and Luquet, P., 1979. Continuousquantitative automatic colletor for fish feces. Prog. FishCult. 41: 64-67.Choubert, G., de la Noue, J. and Luquet, P., 1982. Digestibility infish: improved device for the automatic collection of feces.Aquaculture, 29: 185-189.Christophersen, J., 1968. Effect of freezing and thawing on themicrobial population of foodstuffs. In: Low TemperatureBiology of Foodstuffs (Eds.) J. Hawthorn and E.J. Rolfe,Pergamon Press, New York. pp. 251-269.Chu, C.H. and Bjeldanes, F., 1981. Effect of diamines, polyaminesand tuna fish extracts on the binding of histamine to mucin invitro. J. food Sci. 47: 79-80, 88.184Clancy, G.S., Beames, R.M., Higgs, D.A. and Dosanjh, B.S. 1990.Quality criteria for fish meal: Effect of storage (spoilage)and processing conditions on protein quality. AquacultureInternational Congress and Exposition, September 4-7, 1990,Vancouver, B.C. p. 430.Cladinin, D.R., 1949. The effects of methods of processing on thenutritive value of herring meals. Poult. Sci. 28: 128-133.Cobb, B.F., Alaniz, I. and Thompson, C.A., 1973. Biochemical andmicrobial studies on shrimp: volatile nitrogen and aminonitrogen analysis. J. Food Sci. 38: 431-436.Collie, N.L., 1985. Intestinal nutrient transport in coho salmon(Oncorhynchus kisutch) and the effects of development,starvation and sea water adaptation. J. Comp. Physiol. 156B:163-174.Colnago, G.L. and Jensen, L.S., 1992. 	 Putrescine effects onperformance of male broiler chicks fed low protein dietssupplemented with essential amino acids. Poultry Sci. 71:211-214.Connell, J.J., 1962. The effects of freeze-drying and subsequentstorage on the proteins of flesh foods. In: Freeze-drying ofFoods. (Ed.) F.R. Fisher. Proc. Conf. Freeze-drying, April 12-14, 1961. Chicago, Illinois. pp. 50-58.Connell, J.J., 1975. Control of Fish Quality. Fishing News BooksLtd. Surrey, England. p. 178.Connell, J.J. 1990. Control of Fish Quality. 3rd ed. Fishing NewsBooks Ltd. Surrey, England. p. 227.Tonnell, J.J. and Shewan, J.M. 	 1980. Sensory and non-sensoryassessment of fish. 	 In: 	 Advances in Fish Science andTechnology (Ed.) Connell, J.J.	 Fishing News Books Ltd.,London, pp. 56-64.Conway, E.J. 1957. Microdiffusion Analysis and Volumetric Error.Crosby Lockwood and Son Ltd., London.Conway, E.J. and Byrne, A. 1933. An absorption apparatus for themicro-determination of certain volatile substances. I. themicro determination of ammonia. Biochem. J. 27: 419-429.Cowey, C.B., 1991. The utilization of dietary putrescine by rainbowtrout. Abstract. 20 th Fish feed nutrition workshop, Aug. 14-16,1991, Ithaca, N.Y.Cowey, C.B. and Sargent, J.R., 1979. 	 Nutrition. In: Fish185Physiology. Vol. VII. (Eds.) H.W. Hoar and D.J. Randall.Academic Press, New York. pp. 1-69.Cowey, C.B. and Walton, M.J. 1989. Intermediary metabolism. In:Fish Nutrition 2nd Ed. (Ed) J.E. Halver, Academic Press, NewYork. pp. 260-331.Cowey, C.B., Pope, J.A., Adron, J.W. and Blair, A., 1971. Studieson the nutrition of marine flatfish. Growth of plaice,Pleuronectes platessa, on diets containing proteins derivedfrom plants and other sources. Mar. Biol., 10: 145-153.Cowey, C.B., Degener, E., Tacon, A.G.J., Youngson, A. and Bell,J.G., 1984. The effect of vitamin E and oxidized fish oil onthe nutrition of rainbow trout (Salmo gairdneri) grown atnatural, varying water temperatures. Brit. J. Nutr., 51:443-451.Crowder, B. 1990. U.S. demand for fish meal. In: Making Profitsout of Seafood Wastes. Proc. Int. Conf. on Fish-Byproducts,Anchorage, Alaska April 25-27. p. 57-60.Cui, Y. and Liu, J. 1990. Comparison of energy budgets amoung sixteleosts II. Metabolic rates. Comp. Biochem. Physiol. 97A:169-174.Dabrowski, K., 1983. Comparative aspects of protein digestion andamino acid absorption in fish and other animals. Comp.Biochem. Physiol. 74A: 417-426.Dabrowski, K. and Dabrowska, H., 1981. Digestion of protein byrainbow trout (Salmo gairdneri Rich.) and absorption of aminoacids within the alimentary tract. Comp. Biochem. Physiol.69A: 99-111.Dabrowski, K., Leray, C., Nonnotte, G. and Colin, D.A. 1986.Protein digestion and ion concentrations in rainbow trout(Salmo gairdneri Richardson) digestive tract in sea and freshwater. Comp. Biochem. Physiol. 83A: 27-39.Daikoku, T., Arai, S., Murata, M. and Sakaguchi, M., 1987. Effectsof dietary trimethylamine on the concentration oftrimethylamine oxide in various tissues of the chum salmon,Oncorhyncus keta, in freshwater and seawater environments.Comp. Biochem. Physiol., 87A: 101-105.Damoglou, A.P. 1980. A comparison of different methods offreshness assessment of herring. In: Advances in Fish Siariceand Technology. (Ed.) J.J. Connell, p. 394-399. Fishing NewsBooks Ltd. Surrey, England.186De la Noue, J. and Choubert, G., 1986. Digestibility in rainbowtrout:comparison of the direct and indirect methods ofmeasurement. Frog. Fish. Cult., 48: 190-195.Delong, D.C., Halver, J.E. and Mertz, E.T. 1958. The nutrition ofsalmonid fishes VI. Protein requirements of chinook salmon attwo water temperatures. J. Nutr. 65: 589-599.Department of Fisheries and Oceans, 1989. Chemical Methods Manual.Ch. 3- Quality Indices: Putrescine and Cadaverine. 3(3): 1-7.Department of Fisheries and Oceans, 1990. Groundfish stockassessments for the West coast of Canada in 1989 andrecommended yield options for 1990. Can. Tech. Rep. Fish.Aquat. Sci. No. 1732, 343 pp.de Silva, N.N. and Hughes, R.B. 1962. Influence of tetracyclineantibiotics on the spoilage of herring (Clupea harengus). J.Sci. Food Agric. 13: 161-168.de Silva, S.S. and Perera, M.K. 1984. Digestibility inSarotherodon niloticus (Tilapias) fry: effect of dietaryprotein level and salinity with further observations onvariability in daily digestibility. Aquaculture 38: 293-306.de Silva, S.S., Shim, K.F. and Ong-Khim, A. 1990. An evaluation ofthe method used in digestibility estimations of dietaryingredient and comparisons on external and internal markers,and time of faeces collection in digestibility studies in thefish Oreochromis aureus (Steinbachner). Rep. Nutr. Dev. 30:215-226.Douglas, W.W., 1985. Histamine and 5-hydroxytryptamine (serotonin)and their antagonists. Ch. 26. In: The Pharmacological Basisof Therapeutics (Eds.) A. Goodman-Gilman, L.S. Goodman, T.W.Rall and F. Marad. pp. 604-638.Dosanjh, B.S., Higgs, D.A., Plotnikoff, M.D., Markert, J.R. andBuckley, J.T., 1988. Preliminary evaluation of canola oil,pork lard and marine lipid singly and in combination assupplemental dietary lipid sources for juvenile fall chinooksalmon (Oncorhynchus tshawytscha). Aquaculture, 68: 325-343.DPA Group Inc. 1988. Strategy to access critical raw materialsfor fish feed- B.C. Salmon Farming Industry. Report forAquaculture Association of British Columbia.Dyer, W.J. 1945. Amines in fish muscles I. colorimetricdetermination of trimethylamine as the picric acid salt. J.Fish. Res. Bd. Can. 6: 351-358.187Dyer, W.J. 1952. Amines in fish muscle.6. Trimethylamine oxidecontent of fish and marine invertebrates. J. Fish. Res. Bd.Can. 8: 314-324.Dyer, F.E. and Wood, A.J. 1947. Action of enterobacteriaceae oncholine and related compounds. J. Fish. Res. Bd. Can. 7:17-21.Eddy, F.B. 1982. Osmotic and ionic regulation in captive fish withparticular reference to salmonids. Comp. Biochem. Physiol.73B: 125-141.Edmunds, W.J. and Eitenmiller, R.R., 1975. Effect of storage timeand temperature on histamine content and histaminedecarboxylase activity of aquatic species. J. Food Sci. 40:516-519.Eggum, B.O., Hansen, I. and Larsen, T. 1989. Protein quality anddigestible energy of selected foods determined in balancetrials with rats. Plant Food Human Nutr., 39: 13-21.Eid, A.E. and Matty, A.J. 1989. A simple in vitro method formeasuring protein digestibility. Aquaculture 79: 111-119.Eitenmiller, R.R., Orr, J.H. and Wallis, W.W., 1982. Histamineformation in fish: microbiological and biochemical conditions.In: Chemistry and Biochemistry of Marine Food Products (Eds.)R.E. Martin, G.J. Flick, C.E. Hebard and D.R. Ward. AVI Publ.Comp., Westport, Conneticut. pp. 39-50.El-Lakany, S. and March, B.E., 1974. A comparison of chemicalchanges in freeze-dried herring meals and a lipid proteinmodel system. J.Sci. Fd. Agric. 25:889-897.Ellis, R.W. and Smith, R.R., 1984. Determining fat digestibility introut using a metabolic chamber. Prog. Fish. Cult. 46: 116-119.Fange, R. and Grove, D. 1979. Digestion. In: Fish Physiology Vol.3. (Eds) W.S. Hoar, D.J. Randall and J.R. Brett. pp. 162-260.Academic Press, New York.FAO, 1970. Available amino acid content of fish meals. FAO Fish.Rep. No. 92, 66 p.FAO, 1986. The production of fish meal and oil. fish. Tech. PaperNo. 142, Food and Agricultural Organization, Rome, 63 pp.FAO, 1990. Aquaculture production (1985-1988). FAO FisheriesCircular No. 815. Revision 2, Rome, Italy. 136 p.Farber, L. 1965. Freshness tests. In: Fish as Food V.4. (Ed.) G.188Borgstrom. p. 65-125. Academic Press Inc., New York.Farn, G. and Sims, G.G., 1987. Chemical indices of decomposition intuna. In: Seafood Quality Determination (Eds) D.E. Kramer andJ. Liston. Elsevier Science Pubi., Amsterdam. pp. 175-183.Fauconneau, B. 1985. Protein synthesis and protein deposition infish. In: Nutrition and Feeding in Fish. (Eds) C.B. Cowey,A.M. Mackie and J.G. Bell, Academic Press, London. pp. 17-46.Fauconneau, B., Choubert, G., Blanc, D., Breque, J. and Luquet, P.1983. Influence of environmental temperature on flow rate offoodstuffs through the gastrointestinal tract of rainbowtrout. Aquaculture 34: 27-39.Fernandez-Salguero, J. and Mackie, I.M. 1987. Comparative rates ofspoilage of fillets and whole fish during the storage ofhaddock (Melanogrammus aeglefinus) and herring (Clupeaharrengus) as determined by the formation of non-volatileand volatile amines. In. J. Fd. Sci. Tech. 22: 385-390.Ferraris, R.P. and Ahearn, G.A. 1984. Sugar and amino transport infish intestine. Comp. Biochem. Physiol. 77A: 397-413.Ferraris, R.P., Catacutan, M.R., Mabelin, R.L. and Jazul, A.P.1986. Digestibility in milkfish Chanos chanos (Forsskal):effects of protein source, fish size and salinity. Aquaculture59: 93-105.Fiddler, 	 W., 	 Doerr, 	 R.C. 	 and Gates, 	 R.A. 	 1991. 	 Gaschromatographic method for determination of dimethylamine,trimethylamine and trimethylamine oxide in fish-meat frankfurters. J. Assoc. Off. Anal. Chem. 74: 400-403.Finstad, B. and Thomassen, M.S., 1991. Does dietary lipidcomposition affect the osmoregulatory ability of rainbow trout(Oncorhynchus mykiss) at high and low temperatures. Comp.Biochem Physiol. 99A: 463-471.Flick, G.J., Enriquez, L.G. and Hubbard, J.B. 1986. Shelf life offish and shellfish. In: Handbook of Food and BeverageStability. (Ed.) G. Charalambous. p. 113-351. Academic PressInc., New York.Foltz, J.W., 1979. Comments: Chromic oxide in food assimilationstudies. Trans. Amer. Fish. Soc. 108: 650-651.Fosbol, P. 1985. The effect of careful drying and processing in thequality of fish meal. International Assoc. Of Fish MealManufacturers 25 th Annual Conf., Atlas, Denmark. p. 28-49.189Forster, R.P. and Goldstein, L. 1969. Formation of excretoryproducts. In: Fish Physiology Vol. 1. (Eds) W.S. Hoar and D.J.Randall. pp. 313-350. Academic Press, New York.Frank, H.A., 1985. Histamine forming bacteria in tuna and othermarine fish. In: Histamine in Marine Products: Production byBacteria, Measurement and Prediction of Formation. (Eds) B.S.Pan and D. James. FAO Fish Tech Paper #252: 2-3.Frank, H.A. and Yoshinaga, D.H., 1987. Table for estimatinghistamine formation in skipjack tuna, Katsuwonus pelamis, atlow non-freezing temperatures. Mar. Fish Rev. 49: 67-70.Furuichi, M. and Yone, Y. 1982. Effect of insulin on blood sugarlevels of fishes. Bull. Jap. Soc. Sci. Fish. 48: 1289-1291.Gallardo, J.M., Perez-Martin, R.I., Franco, S.A. and Sotelo, C.G.,1990. Changes in volatile base and trimethylamine oxide duringcanning of albacore (Thunnus alalunga). Int. J. Food Sci.Tech. 25: 78-81.Gardner, H.W., 1979. Lipid hydroperoxide reactivity with proteinsand amino acids: a review. J. Agric. Food Chem. 27: 220-229.Gardner, M.L.G. 	 1988. 	 Gastrointestinal absorption of intactproteins. Ann. Rev. Nutr. 8: 329-350.Gaugas, J.M., 1980. Polyamines in Biomedical Research. Wiley andSons, New York.Gibson, D.M. 1982. A note on the estimation of trimethylamine infish muscle. Fish. Bull. 80: 157.Gildberg, A. 1988. Aspartic proteinases in fishes and aquaticinvertebrates: a review. Comp. Biochem. Physiol. 91B: 425-435.Gill, T.A. and Thompson, J.W. 1984. The rapid, automated analysesof amines in seafood by ion-moderated partition HPLC. J. FoodSci. 49: 603-608.Goldstein, L. 1982. Gill nitrogen excretion. In: Gills (Eds) D.F.Houlihan, J.C. Rankin and T.J. Shuttleworth. CambridgeUniversity Press, New York. pp. 193-206.Goldstein, L. and Forster, R.P. 1970. Nitrogen metabolism infishes. In: Comparative Biochemistry of Nitrogen Metabolism,Vol. II. The Vertebrates. (Ed) J.W. Cambell. Academic Press,New York. pp. 495-518.Gould, E. and Peters, J.A. 1971. On Testing the Freshness ofFrozen Fish. Fishing News Books Ltd., London, England.190Gouygou, J.P., Sinquin, C. and Durand, P., 1987. High pressureliquid chromatography determination of histamine in fish. J.Food Sci. 52: 925-927.Govoni, J.J., Boehlert, G.W. and Watanabe, Y. 1986. Thephysiology of digestion in fish larvae. Environ. Biol. Fish.16: 59-77.Grabner, M., 1985. An in vitro method for measuring proteindigestibility of fish feed components. Aquaculture 48: 97-110.Grabner, M. and Hofer, R., 1985. The digestibility of the proteinsof broad bean (Vicia faba) and soyabean (Glycine max) under invitro conditions simulating the alimentary tracts of rainbowtrout (Salmo gairdneri) and carp (Cyprinus carpio).Aquaculture 48: 111-112.Grant, J. 1989. Fish meal: what is it doing to the price of feed.Canadian Aquaculture 5(4): 64-68.Grant, A.L., Holland, R.E., Thomas, J.W., King, K.J. and Liesman,J.S., 1989. Effect of dietary amines on the small intestinein calves fed soybean protein. J. Nutr. 119: 1034-1041.Grant, A.L., Thomas, J.W., King, K.J. and Liesman, J.S., 1990.Effects of dietary amines on small intestine variables inneonatal pigs fed soy protein isolate. J. Anim. Sci., 68:363-371.Greene, D.H.S. and Selivonchick, D.P., 1987. Lipid metabolism infish. Prog. Lipid Res. 26: 53-85.Haard, N.F., 1991. Alternative protein sources. Annual report ofWestern Regional Aquaculture Consortium (WRAC). Report forApril 1, 1990 to March 31, 1991. 16 p.Haard, N.F. 1992. A review of proteolytic enzymes from marineorganisms and their application in the food industry. J.Aquatic Food Prod. Tech. 1: 17-35.Hajen, W.E. 1990. Evaluation of feedstuff digestibility in post-juvenile chinook salmon (Oncorhynchus tshawtscha) in seawater. M.Sc. thesis, Dept. of Animal Science, U.B.C. 196 pp.Hale, M.B. 1972. Making fish protein concentrates by enzymatichydrolysis: A status report on research and some process andproducts studied by NMFs. NOAA Tech. Report NMFs SSRF-657,32 pp.Halland, H. and Njaa, L.R. 1988. Ammonia (NH 3 ) and total volatile191nitrogen (TVN) in preserved and unpreserved stored, wholefish. J. Sci. Food. Agric. 44: 335-342.Halland, H. and Njaa, L.R., 1989. Total volatile nitrogen - aquality criterion for fish silage? Aquaculture 79:311-316.Halland, H., Arnesen, E. and Njaa, L.R. 1988. Technical note:amino acid composition of whole herring over the spawningseason and during anaerobic storage at 2°C. Int. J. FoodSci. Tech. 23: 521-524.Halland, H., Arenesen, E. and Njaa, L.R. 1990. Amino acidcomposition of whole mackerel (Scromber scrombrus) storedanaerobically at 20 and 2°C. Int. J. Food Sci. Tech. 25:82-87.Hardy, R., 1980. Fish lipids. Part 2. In: Advances in Fish Scienceand Technology. (Ed.) J.J. Connell, Fishing News Books,England. pp. 103-111.Hardy, R.W. 	 1987. 	 Fish silage and liquified fish products.Aquaculture Magazine 2:48-50.Hardy, R.W. and Masumoto, T. 1990. Specifications for marine by-products for aquaculture. p. 109-120. In: Making Profits outof Seafood Wastes: Proc. Int. Conf. on Fish By-Products, April25-27, 1990, Anchorage, Alaska.Hardy, R. and Smith, J.G.M., 1976. The storage of mackerel (Scomberscomber). Development of histamine and rancidity. J. Sci. FoodAgric. 27: 595-599.Hashimoto, Y. and Okaichi, T. 1957. On the determination oftrimethylamine and trimethylamine oxide. A modification ofthe Dyer method. Bull. Jap. Soc. Scient. Fish. 23: 269-272.Hastings, W.H. 1969. Nutritional score. In: Fish in Research.(Eds) O.W. Neuhaus and J.E. Halver. Academic Press, New York.pp. 263-292.Hata, K. and Kaneda, T., 1980. Effect of autoxidized oil on carp.Bull. Jap. Soc. Sci. Fish. 46: 997-1000.Hata, K., Fujimoto, K. and Kaneda, T., 1986. Absorption of lipidhydroperoxides in carp. Bull. Jap. Soc. Sci. Fish. 52: 677-684.Hebard, C.E., Flick, G.J. and Martin, R.E. 1982. Occurrence andsignificance of trimethylamine oxide and its derivatives infish and shellfish. In: Chemistry and Biochemistry of MarineFood Products. (Ed.) R.E. Martin, G.J. Flick and D.R. Ward.p. 149-272. AVI Publishing Co., Westport, CT.192Higgs, D.A. 1986. Feeding fish efficiently. Canadian Aquaculture2(2): 	 15-18, 26.Higgs, D.A., McBride, J.R., Dosanjh, B.S. and Fargerlund, U.H.M.1990. Potential for using canola meal and oil in fish diets.In: 	 Fish Physiology, fish toxicology and fisheriesmanagement: 	 Proceedings of an international symposium,Guangzhou, PRC, September 14-16, 1988. (Ed.) 	 Ryans, R.C.,pp. 88-107.Hilton, J.W. and Slinger, S.J. 1981. Nutrition and feeding ofrainbow trout. Can. Spec. Publ. Fish. Aquat. Sci. Vol. 55, 15Hjorth-Hansen, S. and Bakken, K. 1947. Investigations onanalytical methods for estimation of ammonia and irethylarninesin fish. Reports from the Norwegian Fisheries . Laboratory1(6):1-65.Hobbs, G. 1987. Microbiology of fish. In: Essays in Agriculturaland Food Microbiology. (Eds) J.R. Norris and G.L. Pettipher.p. 119-226. John Wiley and Sons Ltd., Toronto.Hobbs, G. and Hodgkiss, W. 1982. The bacteriology of fish handlingand processing. In: Developments in Food Microbiology. 1 . (Ed.)R. Davies p. 71-117. Applied Science Publishing, London.Hochachka, P.W. 1969. Intermediary metabolism in fishes. FishPhysiology Vol. 1. (Eds) W.S. Hoar and D.J. Randall. pp. 351-390. Academic Press, New York.Hofer, R. 1982. Protein digestion and proteolytic activity in thedigestive tract of an omnivorous cyprinid. Comp. Biochem.Physiol. 72A: 55-63.Hollingworth, T.A., Wekell, M.M., Sullivan, J.J., Torkelson, J.D.and Throm, H.R., 1990. Chemical indicators of decompositionfor raw surimi and flaked artificial crab. J. Food Sci. 55:349-352.Hosomi, M., Stace, N.H., Lirussi, F., Smith, S.M., Murphy, G.M. andDowling, R.H., 1987. Role of polyamines in intestinaladaptation in the rat. Europ. J. Clin. Inv. 17: 375-385.Hossain, H.A. and Jauncey, K., 1989. Studies on the protein, energyand amino acid digestibility of fish meal, mustard oilcake,linseed and sesame oil for common carp (Cyprinus carpio L.).Aquaculture 83: 59-72.Hoyland, D.V. and Taylor, A.J., 1991. A review of the methodologyof the 2-thiobarbituric acid test. Food Chem., 40: 271-279.Hsieh, R.J. and Kinsella, J.E., 1989. Oxidation of polyunsaturated193fatty acids: mechanisms, products and inhibition with emphasison fish. Adv. Food Nutr. Res. 33: 233-341.Hsu, H.W., Vavak, D.L., Satterlee, L.D. and Miller, G.A. 1977. Amultienzyme technique for estimating protein digestibility. J.Food Sci. 42: 1269-1273.Hughes, R.B. 1958. Volatile amines of herring flesh. Nature 181:1281-1282.Hughes, R.B. 1959. Chemical studies on the herring (Clupeaharengus) Trimethylamine oxide and volatile amines in fresh,spoiling and cooked herring flesh. J. Sci. Food Agric. 10:431-436.Hughes, S.G., 1991. Response of first feeding spring chinook salmonto four potential chemical modifiers of feed intake. Prog.Fish. Cult. 53: 15-17.Hughes, S.G. and Barrows, R. 1990. Measurements of the abilitiesof cultured fishes to moisturize their digesta. Comp. Biochem.Physiol. 96A: 109-111.Hui, J. Y . and Taylor, S . L. , 1983. 	 High pressure liquidchromatographic determination of putrefactive amines in foods.J. A. 0. A. C. 66: 853-857.Hung, S.S.O., Cho, C.Y. and Slinger, S.J., 1981. Effect of oxidizedfish oil, DL-alpha-tocopheryl acetate and ethoxyquinsupplementation on vitamin E nutrition of rainbow trout (Salmogairdneri) fed pratical diets. J. Nutr. 111: 648-657.Hungerford, J.M., Walker, K.D., Wekell, M.M., La Rose, J.E. andThrom, H.R., 1990. Selection dtermination of histamine by flowinjection analysis. Anal. Chem. 62: 1971-1976.Huss, H.H. 1988. Fresh fish-quality and quality changes. FAO FishSeries No. 29. p.127.Iijima, N., Zama, K. and Kayama, M., 1983. Effect of oxidizedlipids on the metabolic pathway of lipid biosynthesis in theintestine of carp. Bull. Jap. Soc. Sci. Fish. 49: 1465-1470.Inaba, D., Ogino, C., Takamatsu, C., Sugano, S., and Hata, H.,1962. Digestibility of dietary components in fishes. 1.digestibility of dietary proteins in rainbow trout. Bul. Jap.Soc. Sci. Fish. 28: 367-371.Ishida, Y., Fujii, T. and Kadota, H., 1976. Microbiological studieson salted fish stored at low temperature: chemical andmicrobial changes of salted fish during storage. Bull. Jap.Soc. Sci. Fish. 42: 351-358.194Ito, Y., Terao, H., Noguchi, T. and Naito, H., 1988. Gizzerosineraises the intracellular cyclic adenosine-3',5'-monophosphatelevel in isolated chicken proventriculus. Poultry Sci. 67:1290-1294.Jacobsen, F., 1985. Effect of enzymatic treatment of stickwater onevaporation capacity and fouling. Process Biochem 19: 103-108.Jangaard, P.M., 1987. Fish Silage: A review and some recentdevelopments. In: Proc. Fish Silage Workshop, Univ. St. Anne,Church Pt., N.S. June 17, 1987. Gen. Ed. Ser. #7, Fish Dev.Branch, DFO. pp. 8-33.Jangaard, P.M., 1991. Fish silage in Norway: current activities.Bull. Aquaculture Assoc. Canada. 91(1): 28-53.Jiang, S.T., Hwang, D.C. and Chen, C.S., 1988. Denaturation andchange in SH group of actomyosin from milkfish (Chanos chanos)during storage at -20°C. J. Agric. Food Chem. 36: 433-437.Jobling, M. 1983. A short review and critique of methodology usedin fish growth and nutrition studies. J. Fish Biol. 23: 685-703.Johnston, J. and Coon, C.N., 1979. A comparison of six proteinquality assays using commercially available protein meals.Poultry Sci. 58: 919-927.Jones, K.A., 1990. Chemical requirements of feeding in rainbowtrout, Oncorhyncus mykiss (W); palatability studies on aminoacids, amides, amines, alcohols, aldehydes, saccharides andother compounds. J. Fish. Biol. 37: 413-423.Jung, H.Y.P. and Bjeldanes, L.F. 1979. Effects of cadaverine onhistamine transport and metabolism in isolated gut sections ofthe guinea-pig. Fd. Cosmet. Toxicol. 17: 629-632.Jurss, K., Bittorf, T.H., Vokler, T.H. and Wacke, R. 1983.Influence of nutrition on biochemical sea water adaptation ofthe rainbow trout (Salmo gairdneri Richardson) Comp. Biochem.Physiol. 75B: 713-717.Kapoor, B.G., Smit, H. and Verighina, I.A. 1975. The alimentarycanal and digestion in teleosts. Adv. Mar. Biol. 13: 109-239.Karmas, E. and Mietz, J.L., 1978. Polyamine and histamine contentof tuna fish and the relationship to decomposition. LebensmWiss U. Technol. 11: 333-337.Kaushik, S.J. 1989. Nitrogen and energy balance studies in fish.Proc. of the Tagung zur Verleihung des Roche Research Prize195for Animal Nutrition, Donnerstag, November 9, 1989. pp. 71-85.Kaushik, S.J., Medale, F., Fauconneau, B. and Blanc, D., 1989.Effect of digestible carbohydrates on protein/energyutilization and on glucose metabolism in rainbow trout (Salmogairdneri R.). Aquaculture 79: 63-74.Kawabata, T., Ishizaka, K. and Miura, T., 1956. Studies on theallergy like food poisoning associated with putrefaction ofmarine products. VI. Influence of trimethylamine andtrimethylamine oxide added to histamine or " saurine " upontheir own action in vivo. Nippon Suisan Gakkaishi 21: 1177-1180.Keay, J.N. and Hardy, R. 1972. The separation of aliphatic aminesin dilute aqueous solution by gas chromatography andapplication to the quantitative analysis of tri anddimethylamine in fish. J. Sci. Fd. Agric. 23: 9-19.Ketola, H.G., Smith, C.E. and Kinkschi, G.A. 1989. Influence ofdiet and oxidative rancidity on fry of Atlantic and cohosalmon. Aquaculture 79:417-424.Khayat, A. and Schwall, D., 1983. Lipid oxidation in seafood. FoodTech. 37(10): 130-140.Kilpatrick, J.S. 1990. Trends for the future. In: Making Profitsout of Seafood Wastes. Proc. Int. Conf. on Fish Byproducts,April 25-27, 1990. Anchorage, Alaska. p. 225-230.Kim, K.I., Kayes, T.B. and Amundson, C.H., 1991. Purified dietdevelopment and re-evaluation of the dietary proteinrequirement of fingerling rainbow trout (Oncorhyncus mykiss).Aquaculture 96: 57-67.Kimata, M., 1961. The histamine problem. In: Fish as Food. Vol. 1.(Ed) G. Borgstrom. Academic Press, New York. pp. 329-352.Kitamikado, M. and Tachino, S., 1960. Studies on the digestiveenzymes of rainbow trout. II. Proteases. Bull. Jpn. Soc. Sci.Fish. 26: 685-690.Klausen, N.K. and Lund, E., 1986. Formation of biogenic amines inherring and mackerel. Z. Lebenesm. Unters Forsch 182: 459-463.Koizumi, C., Ohshima, T. and Wada, S., 1990. A comparison of thedeterioration in quality of freeze dried flesh of halibut(Hippoglossus stenolpis) during storage in air and nitrogen atlow water activities. J. Sci. Food Agric. 52: 561-572.Kolb, A.K. and Luckey, T.D., 1972. Markers in nutrition. Nutr.196Abstr. Rev. 42: 813-845.Kramer, D.E., Nordin, D.M.A. and Gardner, L.J., 1977. A comparisonof the quality changes of Alaska pollock and Pacific codduring frozen storage at -28°C. Fish Mar. Ser. Tech. Rep. No.753.Kanazawa, A., 1985. Essential fatty acid and lipid requirement offish. In: Nutrition and Feeding in Fish. (Eds.) C.B. Cowey,A.M. Mackie and J.G. Bell. Academic Press, London. pp. 281-298.Kreiberg, H. 1991. Effect of ration level and water temperature ongrowth and conversion efficiency in chinook salmon inseawater. World Aquaculture 22(1): 84-85.Kristjansson, M.M. 1991. 	 Purification and characterization oftrypsin from the pyloric caeca of rainbow trout (Oncorhynchusmykiss) J. Agric. Food Chem. 39: 1738-1742.Krzymien, M.E. and Elias, L. 1990. Feasibility study on thedetermination of fish freshness by trimethylamine headspaceanalysis. J. Food Sci. 55: 1228-1232.Lall, S.P. 	 1989. 	 The minerals. 	 In: 	 Fish Nutrition, SecondEdition. (Ed.) Halver, J.E. Academic Press, London, pp. 219-257.Lall, S.P. and Bishop, F.J. 	 1979. 	 Studies on the nutrientrequirements of rainbow trout, Salmo gairdneri, grown inseawater and freshwater. 	 In: 	 Advances in Aquaculture.(Eds.) 	 Pilley, I.V. and Dill, D.A. 	 Fishing News Books,Farnham, Great Britain. pp. 580-584.Lassen, S., Bacon, E.K. and Dunn, H.J. 1949. The relationship ofnutritive value of condensed fish solubles to quality of rawmaterial. Poultry Sci. 28:134-140.Lassen, S., Bacon, E.K. and Dunn, H.J. 	 1951. 	 Fish reductionprocess: Relationship of yields and quality of products tofreshness of raw material. Ind. Eng. Chem. 43:2082-2087.Lathia, D.N. and Koch, M., 1989. Comparative study of phytic acidcontent, in-vitro protein digestibility and amino acidcomposition of different types of flat breads. J. Sci. FoodAgric. 47: 353-364.Lebiedzinska, A., Lamparczyk, H., Ganowiak, Z. and Eller, K., 1991.Differences in biogenic amine patterns in fish obtained fromcommercial sources. Z. Lebensm. Unters Forsch 192: 240-243.197Leger, C., 1985. Digestion, absorption and transport of lipids.In: Nutrition and Feeding in Fish. (Eds.) C.B. Cowey, A.M.Mackie and J.G. Bell. Academic Press, London. pp. 299-331.Lerke, P.A., Porcuma, M.N. and Chin, H.B., 1983. Screening test forhistamine in fish. J. Food Sci. 48: 155-157.Licciardello, J.J. and D'Entremont, D.L., 1987. Bacterial growthrate in iced fresh or frozen thawed Atlantic cod, Gadusmorhua. Mar. Fish. Rev. 49: 43-45.Lied, E., Julshamn, K. and Braekkan, O.R., 1982. Determination ofprotein digestibility in Atlantic cod (Gadus morhua) withinternal and external indicators. Can. J. Fish. Aquat. Sci.39: 854-861.Liston, J. 1980. Microbiology in fishery sciences. In: Advances inFish Science and Technology. (Ed.) J. Connell. p. 138-157.Fishing News Books Ltd., England.Lovell, T. 1989. Nutrition and Feeding of Fish. AVI Press, NewYork. 255 p.Lovern, J.A., 1965. Some analytical problems in the analysis offish and fish products. J.A.O.A.C. 48:60-68.Lucke, F. and Geidel, W. 1935. Bestimmung des fluchtigenbasischen stickstoffs in fischen als mabstab fur ihrenfrischezustand. Z-Lebensmitt-Untersuch 70: 441-458.Luijpen, A.F.M.G. 1958. Objective spoilage tests for fish storedunder conditions other than normal chilling in ice. J. Sci.Food Agric. 9: 410-417.Mackie, I.M. and Thomson, B.W. 1974. Decomposition oftrimethylamine oxide during iced and frozen storage of wholeand comminuted tissue of fish. Proceedings of the IVInternational Congress of Food Science and Technology, Vol. 1.243-250.MacLeod, M.G. 1977. Effect of salinity on food intake, absorptionand conversion in the rainbow trout Salmo gairdneri. MarineBiol. 43: 93-102.Malle, P. and Tao, S.H. 1987. Rapid quantitative determination oftrimethylamine using steam distillation. J. Fd. Prot. 50:756-760.Malle, P. and Poumeyrol, M. 1989. A new chemical criterion forthe quality control of fish: trimethylamine/total volatilebasic nitrogen (%). J. Fd. Prot. 52: 419-423.198Mann, J.J., 1988, The enzymatic in vitro evaluation of proteinsources for monogastric animals using the pH-stat method. M.Sc. Thesis, University of British Columbia, 134 p.March, B.E., 1962. Fish meal and condensed fish solubles in poultryand livestock feeding. In: Fish as Food, V. II. (Ed.) G.Borgstrom, Academic Press, New York. pp. 377-434.March, B.E. and Hickling, D.R., 1982. Assessment of heat damage toprotein digestibility from fish meals by in vitro pepsinsolubilization at different temperatures. Can. J. Anim. Sci.62: 657-660.March, B.E., Biely, J. and Tarr, H.L.A., 1963. Nutritivecomposition and evaluation of British Columbia whole herringmeal. J. Fish. Res. Bd. Can. 20: 229-238.March, B.E., Biely, J. and Goudie, C., 1966. Protein nutritivequality of commercial fish meals as assessed by biological andchemical methods. J. Fish. Res. Bd. Can. 23: 395-414.March, B.E., Biely, J., Tarr, H.L.A. and Claggett, F.E., 1965. Theeffect of antioxidant treatment on the metabolizable energyand protein value of herring meal. Poult. Sci. 44: 679-685.March, B.E., MacMillan, C. and Ming, F.W., 1985. Techniques forevaluation of dietary protein quality for the rainbow trout(Salmo gairdneri). Aquaculture, 47: 275-292.Marki, B. 1990. Effects of process parameters and raw materialfreshness on fish meal quality. p. 105-108. In: Making Profitsout of Seafood Wastes: Proc. Int. Conf. on Fish By-Products.April 25-27, 1990. Anchorage, Alaska.Masumura, T., Sagahara, M., Noguchi, T., Mori, K. and Naito, H.,1985. The effect of gizzerosine on acid secretion by isolatedmucosal cells of the chicken proventriculus.Matsumoto, J.J., 1979. Denaturation of fish muscle proteins duringfrozen storage. Adv. Chem. Series, V. 180: 205-224.McCallum, I.A. 1985. Qualitative and quantitative aspects of theprotein nutrition of juvenile chinook salmon (Oncorhynchustshawytscha) Ph. D. Thesis, University of British Columbia,199 p.McCallum, I.A. and Higgs, D.A. 1989. An assessment of processingeffects on the nutritive value of marine protein sources forjuvenile chinook salmon (Oncorhynchus tshawytscha).Aquaculture 77:181-200.McLean, E. and Ash, R., 1987. Intact protein (antigen) absorption199in fishes: mechanisms and physiological significance. J. Fish.Biol. 31(Supplement A): 219-223McLean, E. and Ash, R., 1989. Chronic cannulation of the hepaticportal vein in rainbow trout, Salmo gairdneri: a prerequisiteto not absorption studies. Aquaculture 78: 195-205.McLean, E. and Donaldson, E.M., 1990. Absorption of bioactiveproteins by the gastrointestinal tract of fish: A review. J.Aquatic Anim. Health 2: 1-11.McLean, E., Von Der Meden, A.C. and Donaldson, E.M. 1990. Directand indirect evidence for polypeptide absorption by theteleost gastrointestinal tract. J. Fish. Biol. 36: 489-498.Mearns, K.J., Ellingsen, 0.F., Doving, K.B. and Helmer, S., 1987.Feeding behavior in adult rainbow trout and Atlantic salmonparr, elicited by chemical fractions and mixtures of compoundsidentified in shrimp extract. Aquaculture 64: 47-63.Middlebrooks, B.L., Toom, P.M., Douglas, W.L., Harrison, R.E. andMcDowell, S., 1988. Effect of storage time and temperature onthe microflora and amine development in Spanish mackerel(Scomberomorus maculatus). J. Food Sci. 53: 1024-1029.Mietz, J.L. and Karmas, E., 1977. Chemical quality index of cannedtuna as determined by high pressure liquid chromatography. J.Food Sci. 42: 155-158.Mietz, J.L. and Karmas, E., 1978. Polyamine and histamine contentof rockfish, salmon, lobster and shrimp as an indicator ofdecomposition. J. A. 0. A. C. 61: 139-145.Mita, H., Yasueda, H. and Shida, T. 1980. Quantitative analysis ofhistamine in biological samples by gas chromatography-massspectrometry. J. Chrom. 181: 153-159.Mizui, T., Shimono, N. and Doteuchi, M., 1987. A possible mechanismof protection by polyamines against gastric damage induced byacidified ethanol in rats: polyamine protection may depend onits antiperoxidative properties. Japan J. Pharmacol. 44: 43-50.Morii, H., Cann, D.C. and Taylor, L.Y., 1988. Histamine formationby luminous bacteria in mackerel stored at low temperatures.Nipp. Sui. Gak. 54: 299-305.Murai, T., 1992. Protein nutrition of rainbow trout. Aquaculture100: 191-207.Murray, C.K. and Gibson, D.M., 1972a. An investigation of themethod of determining trimethylamine in fish muscle extracts200by the formation of its picrate salt-Part 1. J. FoodTechnol. 7: 35-46.Murray, C.K. and Gibson, D.M., 1972b. An investigation of themethod of determining trimethylamine in fish muscle extractsby the formation of its picrate salt-Part 2. J. Food Technol.7: 47-51.Murray, C.K., Hobbs, G. and Gilbert, R.J., 1982. Scombrotoxin andscombrotoxin-like poisoning from canned fish. J. Hygiene 88:215-221.Molin, G., Stenstrom, I.M. and Ternstrom, A. 1983. The microbialflora of herring fillets after storage in carbon dioxide,nitrogen or air at 2°C. J. Appl. Bact. 55: 49-56.National Research Council, 1981. 	 Nutrient Requirements ofColdwater Fishes. National Research Council, National Academyof Sciences, Washington, D.C., U.S.A. 63 pp.National Research Council, 1983. 	 Nutrient Requirements ofWarmwater Fishes and Shellfishes. National Research Council.National Academy of Sciences, Washington, D.C., U.S.A.102 pp.Njaa, L.R., Opstvedt, J. and Eggum, B.O., 1980. Comparison ofdifferent chemical and biological methods for protein qualityevaluation in one American and nine Scandinavian laboratories.Acta. Agric. Scand., 30: 393-400.Ng, W.K. and Wee, K.L., 1989. The nutritive value of cassava leafmeal in pelleted feed for nile tilapia. Aquaculture 83: 45-58.Nose, T. 1989. Protein and amino acid nutrition in fish. Proc. fromthe Tagung zur Verleihung des Roche Research Prize for AnimalNutrition, Donnerstag, November 9, 1989. pp. 23-50.Obrink, K.J. 1955. A modified conway unit for microdiffusionanalysis. Biochem. J. 59: 134-136.O'Hare, W.T., Currey, M.C. and Allen, J.C., 1984. Effects ofbuffering capacity on a commonly used assay of proteindigestibility. J. Food Sci. 49: 498-499.Ohashi, E., Takao, Y., Takao, F., Shimizu, Y. and Egashira, M.1991. Semiconductive trimethylamine gas sensor for detectingfish freshness. J. Food Sci. 56: 1275-1278.Okazaki, T., Noguchi, T., Igarashi, K., Sakagami, Y., Seto, H.,Mori, K., Naito, H., Masumura, T. and Sugahara, M., 1983.Gizzerosine, a new toxic substance in fish meal causes severeGE in chicks. Agric. Biol. Chem. 47: 2949-2952.201Okuzumi, M., Okuda, S. and Awana, M., 1981. 	 A new type ofhistamine forming bacteria. I. Isolation of psychrophilic andhalophilic histamine forming bacteria from Scomber japonicus.Bull. Jap. Soc. Sci. Fish. 47: 1591-1598.Okuzumi, M., Okuda, S. and Awana, M., 1982. A new type of histamineforming bacteria II. Occurrence of psychrophilic andhalophilic histamine forming bacteria (N-group bacteria) on/inred meat fish. Bull. Jap. Soc. Sci. Fish. 48: 799-804.Okuzumi, M., Awana, M. and Ohki, Y., 1984. Effects of temperature,pH value and NaC1 concentration on histamine formation of N-group bacteria (Psychrophilic and halophilic histamine formingbacteria). Bull. Jap. Soc. Sci. Fish. 50: 1757-1662.Okuzumi, M., Fukumoto, I. and Fujii, T., 1990. 	 Changes inbacterial flora and polyamine contents during storage of horsemackerel meat. Nippon Suisun Gakkaishi 56: 1307-1312.011ey, J. and Pirie, R., 1966. The pepsin digestibility method atlow pepsin strengths. Fish. News Int., 5(12): 27-29.Opstvedt, J., 1973a. Influence of residual lipids on nutritivevalue of fish meal. IV effect of drying and storage on theenergy value of the protein and lipid fractions of herringmeal. Acta Agric. Scand., 23: 200-208.Opstvedt, J., 1973b. Influence of residual lipids on nutritivevalue of fish meal. V. Digestion and deposition of marinefatty acids in chickens. Acta Agric. Scand., 23: 217-224.Opstvedt, J., 1985. Fish lipids in animal nutrition. Int. Assoc.Fish Meal Manuf. Tech. Bull. No. 22. 27 p.Opstvedt, J., 1988. Influence of drying and smoking on proteinquality. In: Fish Smoking and Drying (Ed.) J.R. Burt,Elsevier, New York. p. 23-40.Opstvedt, J., Miller, R., Hardy, R.W. and Spinelli, J. 1984. Heatinduced changes in sulfhydryl groups and disulfide bonds infish protein and their effect on protein and amino aciddigestibility in rainbow trout. J. Agric. Food. Chem. 32:929-935.Otake, S., Maeda, T. and Fukui, K., 1977. Biological investigationon metabolism of histamine in mackerel (Scomber japonicus)Influence of histamine injection into stomach on breakdowm ofthe tissue. Bull. Jap. Soc. Sci. Fish. 43: 477-488.Parker, N.C., 1987. Feed conversion indices: Controversy orconvention? Prog. Fish Cult. 49: 161-166.202Pan, B.S., 1985. Histamine formation in canning process. In:Histamine in marine products: production by bacteria,measurement and prediction of formation. (Eds.) B.S. Pan andD. James. FAO Fish Tech. Paper No. 252. pp. 41-44.Pan, B.S., 1988. Undesirable factors in dried fish products. In:Fish Smoking and Drying on the Nutritional Properties of Fish.(Ed) J.R. Burt. Elsevier Applied Science, London. pp. 61-72.Pearson, D. and Muslemuddin, M. 1968. The accurate determinationof total volatile nitrogen in meat and fish. 1. Techniquesand application to beef and salmon. J. Assoc. PublicAnalysts 6: 117-123.Pearson, D. and Muslemuddin, M. 1969. The accurate determinationof total volatile nitrogen in meat and fish. 2. Application towhite teleostean fish. J. Assoc. Public Analysts 7: 50-54.Pedersen, B. and Eggum, B.O., 1983. Prediction of proteindigestibility by an in vitro enzymatic pH stat procedure. Z.f Tierphysiol Tierernahrg u Futter mittelkde 49: 265-277.Pedersen, B.H. and Hjelmeland, K. 1988. Fate of trypsin andassimilation efficiency in larval herring (Clupea harrengus)following the digestion of copepods. Marine Biology 97: 467-476.Pedrosa-Menabrito, A. and Regenstein, J.M. 1988. Shelf lifeextension of fresh fish-A review. Part 1-Spoilage of fish. J.Food Quality 11: 117-127.Pegg, A.E. and McCann, P.P., 1982. Polyamine metabolism andfunction. Am. J. Physiol. 243 (Cell Physiol. 12): C212-C221.Perez-Martin, R.I., Franco, J.M., Molist, P. and Gallardo, J.M.1987. Gas chromatographic methods for the determination ofvolatile amines in seafoods. Int. J. Fd. Sci. Tech. 22:509-514.Pfeffer, E., 1982. Utilization of dietary protein by salmonidfish. Comp. Biochem. Physiol. 73B: 51-57.Phillips, R.D. and Finley, J.W., 1990. Protein Quality and theEffects of Processing. Marcel Dekker Inc., New York. 389 p.Pike, I.H., 1990. The role of fish oil in needs for farmed fish.Int. Assoc. Fish Meal Manuf. Tech. Bull. No. 25. 12 p.Pike, I.H., Andorsdottir, G. and Mundheim, H. 1990. The role offish meal in diets for salmonids. I.A.F.M.M. No. 24 p.35.Pickering, A.D. and Macey, D.J., 1977. Structure, histochemistry203and the effect of handling on the mucous cells of theepidermis of the char Salvelinus alpinus (L). J. Fish Biol.10: 505-512.Plantikow, A. and Plantikow, H., 1985. Alanine aminopeptidase(AAP) activity in the midgut of rainbow trout (Salmo gairdneriR.): the influence of feed quantity and quality, temperatureand osmolarity. Aquaculture 48: 261-276.Power, H.E., Savagaon, K.A., March, B.E. and Biely, J., 1969. Thechemical and nutritive characteristics of herring mealsproduced on the Atlantic coast of Canada. Fish Res. Bd. Can.Tech. Rep. No. 114. 14pPowrie, W.D., 1973. Characteristics of food myosystems and theirbehaviour during freeze-preservation. In: Low TemperaturePreservation of Foods and Living Matter. (Eds.) O.R. Fennema,W.D. Powrie and E.H. Marth. Marcel Dekker Inc., New York.Ch. 6Priestly, R.J., 1979. Effects of Heating on Foodstuffs . AppliedSci. Publ., London. 417 p.Raa, J. and Gildberg, A., 1982. Fish silage: a review. CRC Crit.Rev. Fd. Sci. Nutr. 16: 383-419.Reay, G.A. and Shewan, J.M. 1949. The spoilage of fish and itspreservation by chilling. Adv. Food Res. 2: 344-397.Richardson, N.L., Higgs, D.A., Beames, R.M., and McBride, J.R.1985. Influence of dietary calcium, phosphorous, zinc andsodium phytate level on cataract incidence, growth andhistopathology in juvenile chinook salmon (Oncorhynchustshawytscha). J. Nutr., 115:553-567.Richardson, N.L., Higgs, D.A., and Beames, R.M. 	 1986. 	 Thesusceptibility of juvenile chinook salmon (Oncorhynchustshawytscha) to cataract formation in relation to dietarychanges in early life. Aquaculture 52:237-343.Ringo, E. 1991. Hatchery reared landlocked Arctic char, Salvelinusalpinus (L.) from lake Takvatn reared in fresh and seawater.II. the effect of salinity on the digestibility of protein,lipid and individual fatty acids in a caplin roe diet andcommercial feed.Ritchie, A.H. and Mackie, I.M., 1980. The formation of diamines andpolyamines during storage of mackerel (Scomber scombrus). In:Advances in Fish Science and Technology (Ed.) J.J. Connell, p.489-494. Fishing News Books Ltd., Surrey, England.Ritter, M.A., Morr, C.V. and Thomas, R.L., 1987. In vitro204digestibility of phytate reduced and phenolic reduced soyprotein isolates. J. Food Sci. 52: 325-327, 341.Rosier, J. and Peteghem, C.V., 1988. A screening method for thesimultaneous determination of putrescine, cadaverine,histamine, spermidine and spermine in fish by means of highpressure liquid chromatography of the 5-dimethylaminophthalene-l-sulphonyl derivatives. Z. LebensmUnters Forsch 186: 25-28.Rothenbuhler, E. and Kinsella, J.E., 1985. The pH-stat method forassessing protein digestibility: an evaluation. J. Agric. FoodChem. 33: 433-438.Roubal, W.T. and Tappel, A.L., 1966. Damage to protein, enzymes andamino acids by peroxidizing lipids. Arch. Bioch. Biophys.113: 5-8.Ruiter, A. and Weseman, J.M. 1976. The automated determination ofvolatile bases (trimethylamine, dimethylamine and ammonia) infish and shrimp. J. Fd. Technol. 11: 59-68.Rychly, J. and Spannhof, L., 1979. Nitrogen balance in trout:digestibility of diets containing varying levels of proteinand carbohydrate. Aquaculture 16: 39-46.Saha, D.C. and Gilbreath, R.L., 1991a. Analytical recovery bycolorimetry and atomic absorption spectrophotometry. J. Sci.Food Agric. 55: 433-446.Saha, D.C. and Gilbreath, R.L., 1991b. Faecal recovery and diurnalvariation in excretion of dietary chromium by mature swine. J.Sci Food Agric. 56: 407-418.Sanchez-Chiang, L., Cisternas, E., and Ponce, 0. 1987. Partialpurification of pepsins from adult and juvenile salmon fish(Oncorhynchus keta). Effect of NaC1 on proteolyticactivities. Comp. Biochem. Physiol., 87B:793-797.Sandfeld, P., 1983. 	 Project, quality criteria for fish meal.Assoc. Danish Fish Meal Oil Ind., Copenhagen. 221 pp.Sargent, J., Henderson, R.J. and Tocher, D.R., 1989. The lipids.In: Fish Nutrition, Second Edition. (Ed) J.E. Halver, AcademicPress Inc., New York. pp. 153-219.Satia, B.P., 1974. Quantitative protein requirement of rainbowtrout. Prog. Fish. Cult. 36: 80-85.Schuber, F., 1989. Influence of polyamines on membrane functions:Review article. Biochem. J. 260: 1-10.205Schneider, B.H. and Flatt, W.P., 1975. The evaluation of feedsthrough digestibility experiments. University of GeorgiaPress, Athens.Seiler, N., 1990. Polyamine metabolism. Digestion 46(Suppl. 2):319-330.Shaw, D.H., Gare, R.L., and Kennedy, M.A. 1977. Chemical andsensory changes during storage of witch flounder(Glyptocephalus cynoglossus) in ice. J. Food Sci. 42: 159-162.Shehadeh, Z.H. and Gordon, M.S. 1969. The role of the intestine insalinity adaptation of the rainbow trout, Salmo gairdneri.Comp. Biochem. Physiol. 30: 397-418.Shenouda, S.Y.K., 1980. Theories of protein denaturation duringfrozen storage of fish flesh. Adv. Food Res. 26: 275-311.Sheridan, M.A., 1988a. Exposure to seawater stimulates lipidmobilization from depot tissues of juvenile coho (Oncorhyncuskisutch) and chinook (0. tshawytscha) salmon. Fish Physiol.Biochem. 5: 173-180.Sheridan, M.A., 1988b. Lipid dynamics in fish: aspects ofabsorption, transport, deposition and mobilization. Comp.Biochem. Physiol. 90B: 679-690.Sheridan, M.A., 1989. Alterations in lipid metabolism accompanyingsmoltification and seawater adaptation of salmonid fish.Aquaculture 82: 191-203.Shewan, J.M. 1961. The microbiology of seawater fish. In: Fish asFood. Vol. 1. (Ed.) G. Borgstrom. p. 487-560. AcademicPress, New York.Shewan, J.M., 1962. The bacteriology of fresh and spoiling fish andsome related chemical changes. In: Recent Advances in FoodScience. (Eds.) J. Hawthorn and J.M. Leitch. p. 167-193.Butterworth and Co. Ltd., London.Shewan, J.M., Gibson, D.M. and Murray, C.K. 1971. The estimationof trimethylamine in fish muscle. In: Fish Inspection andQuality Control. (Ed.) R. Kreuzer. p. 183-186. Fishing NewsBooks Limited, London.Shimakura, K., Yamanaka, H., Shiomi, K. and Kikuchi, T., 1991. Gasliquid chromatographic determination of pyrrolidine andpiperidine in canned fish products. Nipp. Suisun Gakh 57:1753-1757.Sikorski, Z.E. and Naezk, M., 1981. Modification of technological206properties of fish protein concentrates. CRC Crit. Rev. Fd.Sci. Nutr. 14: 201-230.Sikorski, Z.E. and Kolakowska, A., 1990. Freezing of marine foods.In: Seafood Resources, nutritional composition andpreservation. (Ed) Z.E. Sikorski, CRC Press Inc., Boca Raton,Florida. pp. 111-125.Sikorski, Z., 011ey, J. and Kostuch, S. 1976. Protein changes infrozen fish. Crit. Rev. Fd. Sci. Nutr. 8: 97-129.Simpson, B.K., Simpson, M.V. and Haard, N.F. 1990. Properties oftrypsin from the pyloric ceca of Atlantic cod (Gadus morhua).J. Fd. Sci. 55: 959-961, 971.Simpson, B.K., Smith, J.P., Yaylayan, V. and Haard, N.F. 1989.Kinetic and thermodynamic characteristics of a digestiveprotease from Atlantic cod Gadus morhua J. Food Biochem. 13:201-213.Sire, M.F., Lutton, C. and Vernier, J.M. 1981. New views onintestinal absorption of lipids in teleostean fishes: anultrastructural and biochemical study in rainbow trout. J.Lipid Res. 22: 81-94.Slabyj, B.M. and True, R.H., 1978. Effect of preprocessing holdingon the quality of canned Maine sardines. J. Food Sci. 43:1172-1176.Smith, R.R., 1971. A method for measuring digestibility andmetabolizable energy of fish feeds. Prog. Fish Cult. 33: 132-134.Smith, M.V. 1983. Membrane transport in fish intestine. Comp.Biochem. Physiol. 75A: 325-335.Smith, R.R., 1989a. Nutritional energetics. In: Fish Nutrition 2n dEd. (Ed) J.E. Halver. p. 2-31. Academic Press, New York.Smith, L.S. 1989b. Digestive functions in teleost fishes. In: FishNutrition, Second Edition. (Ed) J.E. Halver, Academic PressInc., New York. p. 332-423.Smith, P., 1990a. 	 Innovations in salmon and shrimp feed.Aquaculture International Congress and Exposition, Sept. 4-7,1990, Vancouver, B.C. pp. 121-126.Smith, T.K., 1990b. Effect of dietary putrescine on whole bodygrowth and polyamine metabolism. Proc. Soc. Exp. Biol. Med.194: 332-336.207Smith, B.W. and Lovell, R.T., 1973. 	 Determination of apparentprotein digestibility in feeds for channel catfish. Trans. Am.Fish. Soc. 4: 831-835.Smith, R.R., Rumsey, G.L. and Scott, M.L. 1978. Heat incrementassociated with dietary protein, fat, carbohydrates andcomplete diets in salmonids: comparative energy efficiency. J.Nutr. 108: 1025-1032.Smith, R.R., Pterson, M.C. and Allfred, A.C., 1980. Effect ofleaching on apparent digestion coefficients of feedstuffs forsalmonids. Prog. Fish Cult. 42: 195-199.Smith, G.M., Hardy, R., McDonald, I. and Templeton, J. 1980. Thestorage of herring (Clupea harengus) in ice, refrigerated seawater and at ambient temperatures. Chemical and sensoryassessment. J. Sci. Food Agric. 31: 375-385.Spinelli, J. 1964. Evaluation of the micro-diffusion method forthe determination of tertiary volatile base in marineproducts. NOAA Fish. Ind. Res. 2: 17-19.Spinelli, J. and Koury, B., 1979. Non-enzymic formation ofdimethylamine in dried fishery products. J. Agric. Food Chem.27: 1104-1108.Spyridakis, P., Gabaudan, J., Metailler, R. and Guillaume, J.,1988. Digestibilite des proteines et disponibilite des acidesamines de quelques matieres premieres chez le bar(Dicentrarchus labrax). Reprod. Nutr. Develop. 28: 1509-1517.Spyridakis, P., Metailler, R., Gabaudan, J. and Riaza, A., 1989.Studies on nutrient digestibility in European sea bass(Dicentrarchus labrax); Methodological aspects concerningfaeces collection. Aquaculture 77:61-70.Stansby, M.E., Harrison, R.W., Dassow, J. and Sater, M. 1944.Determining volatile bases in fish. Comparison of precisionof certain methods. Ind. Eng. Chem. 16: 593-595.Staruszkiewicz, W.F., 1977. Fluorometric determination of histaminein tuna: Collaborative study. J. A. 0. A. C. 60: 1131-1136.Staruszkiewicz, W.F. and Bond, J.F., 1981. Gas chromatographicdetermination of cadaverine, putrescine and histamine infoods. J. A. O. A. C. 64: 584-591.Staruszkiewicz, W.F., Waldron, E.M. and Bond, J.F., 1977.Fluorometric determination of histamine in Tuna: developmentof method. J. A. O. A. C. 60: 1125-1130.Steel, G.D. and Torrie, J.H., 1980. Principles and procedures of208statistics: A biometrical approach. 2nd Ed. 633 p.Steffens, W. 1989. Principles of Fish Nutrition. John Wiley andSons, New York. 368 p.Stevenson, A.E. and DeLangen, H. 1960. Measurement of feed intakeby grazing cattle and sheep. N.Z. J. Agric. Res. 5:756-760.Stone, F.E. and Hardy, R.W. 	 1986. 	 Nutritional value of acidstabilised silage and liquefied fish protein. J. Sci. Fd.Agric. 37:797-803.Storebakken, T. and Austreng, E., 1987. Ration level for salmonids.II. Growth, feed intake, protein digestibility, bodycomposition and feed conversion in rainbow trout weighing 0.5-1.0 kg. Aquaculture, 60: 207-221.Swaisgood, H.E. and Catignani, G.L., 1991. Protein digestibility:In vitro methods of assessment. Adv. Food Nutr. Res. 35:185-236.SYSTAT: 	 1990. 	 The system for statistics Version 5.0 (Ed.)Wilkinson, L., Evanston, I.L. SYSTAT Inc. 677 pp.Tabor, C.W. and Tabor, H., 1984. Polyamines. Ann. Rev. Biochem.53: 749-790.Tacon, A.G.J., 1985. Nutritional fish pathology: morphologicalsigns of nutrient deficiency and toxicity in farmed fish. FAOAquaculture and Development and Coordination Programme, ReportNo. 85. 33 p.Tacon, A.G.J. and Cowey, C.B. 1985. Protein and amino acidrequirements. In: Fish energetics: New Perspectives (Eds) P.Tytler and P. Calow. John Hopkins University Press, Baltimore.Tacon, A.G.J. and Rodrigues, A.M.P., 1984. Comparison of chromicoxide, crude fibre, polethylene and acid-insoluble ash asdietary markers for the estimation of apparent digestibilitycoefficients in rainbow trout. Aquaculture, 43: 391-399.Talbot, C., 1985. Laboratory methods in fish feeding andnutritional studies. In: Fish Energetics: New Perspectives(Eds.) P. Tytler and P. Calow. pp. 125-154.Tanaka, M., Nagushima, Y. and Taguchi, T., 1986. A kinetic study ofthe non-enzymatic browning reaction between histamine andglucose. Bull. Jap. Soc. Sci. Fish. 52: 1271-1274.Tarr, H.L.A. 1939. The bacterial reduction of TMAO to TMA. J.Fish. Res. Bd. Can. 4: 367-377.209Tarr, H.L.A. 1940. Specifity of triamineoxidase. J. Fish. Res. Bd.Can. 5: 187-196.Tarr, H.L.A. 1982. Effect of processing on the nutritional valueof fish products used for animal feeding. In: Handbook ofNutritive Value of Processed Feed. (Ed). Rechcigil, M. CRCPress Inc. Boca Raton, Florida. pp. 283-304.Tarr, H.L.A. and Biely, J. 1973. Effect of processing on thenutritional value of fish meal and related products. IN:Effect of Processing on the Nutritional Value of Feeds.National Academy of Sciences, Washington, D.C. pp. 252-281.Tarr, H.L.A., Biely, J. and March, B.E., 1954. The nutritive valueof herring meals. 1. the effect of heat. Poult. Sci. 33: 242-250.Tarr, H.L.A. and Gadd, R.E.A., 1965. Browning of freeze-dried fish.J. Fish Res. Bd. Can. 22: 755-760.Tatterson, I.N. and Windsor, M.L., 1974. Fish silage. TorryAdvisory Note No. 64. 6 p.Taylor, S.L., 1986. Histamine food poisoning: toxicology andclinical aspects. CRC Crit. Rev. Toxic. 17: 191-128.Taylor, S.L. and Lieber, E.R., 1977. Specificity and sensitivity ofseven histamine detection methods. J. Food Sci. 42: 1584-1586.Taylor, S.L. and Lieber, E.R., 1979. In vitro inhibition of ratintestinal histamine metabolizing enzymes. Food Cosmet.Toxicol. 17: 237-240.Taylor, S.L. and Sumner, S.S., 1987. Determination of histamine,putrescine and cadaverine. In: Seafood Quality Determination.(Eds.) D.E. Kramer and J. Liston. Elsevier Science Publ., NewYork. pp. 235-245.Teskeredzic, Z., Teskeredzic, E. and Malmar, L., 1989. Highmortality of rainbow trout (Salmo gairdneri) in Yugoslavianfish farms caused by inadequate feed quality. Aquaculture 79:391-395.Theriault, G. 1989. Fish meal market review. Economic andCommercial Analysis Report No. 28. 88 pp.Thorarensen, H., McLean, E. and Donaldson, E.M. and Farrell, A.P.1991. The blood vasculature of the gastrointestinal tract inchinook salmon (Oncorhyncus tshawytscha) and coho (0. kisutch)salmon. J. Fish Biology 38: 525-531.210Tokunaga, T., Iida, H. and Miwa, K. 1977. The gas chromatographicanalysis of amines in fish. Bull. Jap. Soc. Scient. Fish.43: 219-227.Toyama, K., Okuzumi, M., Yokoi, T. and Aoe, H., 1981. Histaminecontent of fish meal. Bull. Jap. Soc. Sci. Fish. 47: 415-419.Toyama, K., Hoshi, M., Ishiguro, M. and Aoe, H., 1982. Behaviourof histamine during the manufacturing process of fish meal.Bull. Jap. Soc. Fish. 48: 1333-1339.Toyama, K., Hoshi, M., Kamioka, Y., Ohtaka, T., Yamazaki, M. andAoe, H., 1983. Amino acid composition of fish meals and fishsolubles prepared from putrefied raw fish. Bull. Jap. Soc.Sci. Fish. 49: 1221-1231.Toyama, K., Hoshi, M. and Kaizuka, M. 1985. Trigger ability ongizzard erosion by fish meal concerning with a kind of andfreshness of raw fish and manufacturing conditions. Bull.Jap. Soc. Sci. Fish. 51: 985-993.Tozawa, H. and Kawabata, T. 1987. Factors affecting the formationof volatile N-nitrosamines in the preparation process of fishmeals. Nipp. Suis. Gakk. 53: 1449-1456.Tozawa, H., Enokihara, K. and Amano, K. 1971. Proposedmodification	 of 	 Dyer's	 method 	 for 	 trimethylaminedetermination in cod fish. In: Fish Inspection and QualityControl. (Ed.) R. Kreuzer. p. 187-190. Fishing News BooksLtd., London, England.Twining, S.S., Alaxander, P.A., Huibregtse, K. and Glick, D.A.,1983. A pepsinogen from rainbow trout. Comp. Biochem.Physiol. 75B: 109-112.Tyler, A.V. and Fargo, J. 1990. Groundfish stock assessments forthe West coast of Canada in 1989 and recommended yield optionsfor 1990. Can. Tech. Rep. Fish. Aquat. Sci. No. 1732, 343 pp.Tytler, P. and Calow, P. 1985. Fish Energetics: New Perspectives.John Hopkins University Press, Baltimore, Maryland. 349 p.Usher, M.L., Talbot, C. and Eddy, F.B. 1988. Drinking in Atlanticsalmon smolts transferred to seawater and the relationshipbetween drinking and feeding. Aquaculture 73: 237-246.Usher, M.L., Talbot, C. and Eddy, F.B. 1990. Effect of transferto seawater on digestion and gut function in Atlantic salmonsmolts (Salmo salar. L.). Aquaculture 90:85-96.Vens-Cappell, B., 1985. Methodical studies on digestion in trout.1. Reliability of digestion coefficients in relation to211methods for faeces collection. Aquacultural Eng. 4: 33-49.Van Spreekens, K.J.A., 1987. Histamine production by thepsychrophilic flora. In: Seafood Quality Determination (Eds.)D.E. Kramer and J. Liston. Elsevier Sci. Publ., B.V.,Amsterdam. pp. 309-318.Veciana-Nogues, M.T., Vidal-Carou, M.C. and Marine-Font, A., 1990.Histamine and tyramine during storage and spoilage of anchovieEngraulis encrasicholus: relationship with other fish spoilageindicators. J. Food Sci. 55: 1192-1193, 1195.Venugopal, V. 1990. Extracellular proteases of contaminantbacteria in fish spoilage: A review. J. Food Prot. 53(4):341-350.Vyncke, W., Luten, J., Branner, K. and Moermans, R. 1987.Determination of total volatile bases in fish: acollaborative study by the WEFTA. Z. Lebensm Unters Forsch.184: 110-114.Walters, M.J. 1984. Determination of histamine in fish by liquidchromatography with post-column reaction and fluorometricdetection. J. A. 0. A. C. 67: 1041-1043.Walton, M.J. and Cowey, C.B. 	 1982. 	 Aspects of intermediarymetabolism in salmonid fish. Comp. Biochem. Physiol. 73B: 59-80.Ward, A., Wignall, J. and Windsor, M.L. 1977.	 The effect ofcooking temperatures and applied pressure on the release ofliquor from the fish. J. Sci. Fd. Agric. 28: 941-947.Watanabe, T., 1982. Lipid nutrition in fish. Comp. Biochem.Physiol. 73B: 3-15.Watanabe, T. and Pongmanerat, J., 1991. Quality evaluation of someanimal protein sources for rainbow trout (Oncorhynchusmykiss). Nippon Suisan Gakkaishi 57: 495-501.Watanabe, T., Nanri, H., Satoh, S., Takeuchi, M. and Nose, T.,1983. Nutritional evaluation of brown meals as a prtein sourcein diets for rainbow trout. Bull. Jap. Soc. Sci. Fish. 49:1083-1087.Watanabe, T., Takeuchi, T., Satoh, S., Toyama, K. and Okuzumi, M.1987. Effect of dietary histidine on growth and stomacherosion in rainbow trout. Nip. Suis. Gakk. 53: 1207-1214.Watanabe, E., Mogami, K., Hayashi, T. and Toyama, K., 1991. A trialof quality evaluation of fish meal manufactured from sardineby measuring the B-buffer capacity of the water extract. Nip.212Suis. Gak. 57: 951-955.Watson, D.W. 1939. Studies of fish spoilage IV. The bacterialreduction of trimethylamine oxide. J. Fish. Res. Bd. Can. 4:252-266.Wendakoon, C.N., Murata, M. and Sakaguchi, M., 1990. Comparison ofnon-volatile amine formation between the dark and whitemuscles of mackerel during storage. Nip. Suis. Gak. 56: 809-818.Wessels, J.P.H. and Post, B.J. 1989. Effect of heat treatment offish meals, fines and the addition of lysine as related togizzard erosion in chickens. J. Sci. Food Agric. 46: 393-406.Wilson, R.P., 1985. 	 Amino acid and protein requirements. In:Nutrition and Feeding in Fish. (Eds) C.B. Cowey, A.M. Mackieand J.G. Bell. Academic Press, New York. pp. 1-16.Wilson, R.P., 1989. 	 Amino acids and proteins. 	 In: 	 FishNutrition, Second Edition. 	 (Ed.) J.E. Halver, 	 AcademicPress, New york. pp. 112-153.Windell, J.T., Foltz, J.W. and Sarokon, J.A. 1978a. Effect of fishsize, temperature and amount fed on nutrient digestibility ofa pelleted diet by rainbow trout, Salmo gairdneri. Trans. Am.Fish. Soc. 107: 613-616.Windell, J.T., Foltz, J.W. and Sarokon, J.A., 1978b. Methods offecal collection and nutrient leaching in digestibilitystudies. Prog. Fish Cult. 40: 51-55.Windsor, M.L., 1969. Fish protein concentrate. Torry Advisory NoteNo. 39. 8 p.Windsor, M.L., 1971. Fishmeal. Torry Advisory Note No. 49. 12 p.Windsor, M.L. and Barlow, S., 1981. Introduction to Fishery By-products. Fishing News Books Ltd., England. 187 p.Wiseman, J., 1986. Antinutritional factors associated with dietaryfats and oils. In: Recent Advances in Animal Nutrition. (Eds.)W. Haresign and D.J.A. Cole. Butterworths, London. pp. 47-75.Wong, K. and Gill, T.A. 1987. Enzymatic determination oftrimethylamine and its relationship to fish quality. J. FoodSci. 52: 1-3.Wong, K, Bartlett, F. and Gill, T.A. 1988. A diagnostic teststrip for the semiquantitative determination oftrimethylamine in fish. J. Food Sci. 53: 1653-1655.213Woyewoda, A.D., Shaw, S.J., Ke, P.J. and Burns, B.G. 1986.Recommended laboratory methods for the assessment of fishquality. Can. Tech. Rep. Fish. Aquat. Sci. No. 1448 p. 143.Yamanaka, H., Shiomi, K., Kikuchi, T. and Okuzumi, M. 1984. Changesis histamine content in red meat fish during storage atdifferent temperatures. Bull. Jap. Soc. Sci. Fish. 50: 695-701.Yamanaka, H., Shimakura, K. and Shiomi, K. and Kikuchi, T. 1986.Changes in non-volatile amine content of meats and sardinesand saury pike during storage. Bull. Jap. Soc. Sci. Fish. 52:127-130.Yamanaka, H., Shiomi, K. and Kikuchi, T. 1989. Cadaverine as apotential index for decomposition of salmonid fishes. J. FoodHyg. Soc. Japan 30: 170-174.Yen, G.C. and Hsieh, C.L., 1991. Simultaneous analysis of biogenicamines in canned fish by HPLC. J. Food Sci. 56: 158-160.Zeitoun, I.H., Halver, J.E., Ullrey, D.E. and Tack, P.I. 1973.Influence of salinity on protein requirements of rainbow trout(Salmo gairdneri) fingerlings. J. Fish. Res. Bd. Can. 30:1867-1873.Zeitoun, I.H., Ullrey, D.E., Halver, J.E., Tack, P.I. and Magee,W.T. 1974. Influence of salinity on protein requirements ofcoho salmon (Oncorhncus kisutch) smolts. J. Fish. Res. Bd.Can. 31: 1145-1148.214APPENDIX 1Table 1. Comparison of chromic oxide values (mg/g) for Rainbowtrout feces (herring meal study) subjected to grinding (mortor andpestle vs electric coffee grinder) and screening(250 um sieve)prior to analysis. Values in parenthesis are given as the mean ±SD.TANK #405420403407MORTOR AND PESTLE COFFEE GRINDER Level of pa b ab	2.14,	 2.29, 	 2.15	(2.19 	 + 	 .08)2.25, 	 2.38, 	 2.08,2.21 	 (2.23 	 + 	 .12)ns2.32, 	 2.20, 	 1.97 2.18, 	 2.20, 	 2.08 ns(2.16 ±	 .18) (2.15 ±	 .06)2.11, 	 2.31,	 2.22, 2.12, 	 2.23, 	 2.26, ns2.19, 	 2.15 2.03 	 (2.16 	 + 	 .11)(2.20 ±	 .07)NOT SCREENED SCREENED Level of pa b ab2.32, 	 1.77, 	 2.39 1.84, 	 2.22, 	 2.14 ns(2.18 	 +	 .35) (2.07 ±	 .20)	2.4 ,	 2.23, 	 2.70	(2.44 	 + 	 .24)	2.57, 	 2.87, 	 1.94,	2.37, 	 2.61, 	 2.41,ns(2.42 ±	 .31)2.25, 	 1.87, 	 1.86, 2.22, 	 2.23, 	 2.06 ns1.93, 	 2.00 (2.17	 + 	 .10)(1.97 ± .14)400401Statistical differences were tested with paired t-test (SYSTAT,ns p>0.05).215Table 2. Recovery of chromic oxide from a feed mash and Rainbowtrout feces using the wet ashing colorimetric method of Stevensonand De Langen (1960).WEIGHT OF WEIGHT OF 	 EXPECTED 	 ASSAYED 	 RANGE %SAMPLE (g) 	 Cr2O3 (g) 	 Cr2O3 	 Cr2O3 	 RECOVERY	(mg/ )	 (mg/g)FECES34.00 .1526 4.47 4.34 ±.03 91.4-102.834.01 .4362 12.66 12.61 ±.09 92.7-106.334.01 .4953 14.35 13.82 +.10 89.3-103.3FEED MASH509.2 3.25 6.34 6.39 ±.01 99.2-102.4511.6 6.02 11.63 11.66 ±.02 98.5-102.0499.0 8.52 16.77 16.74 ±.02 98.6-101.0511.8 11.20 21.42 21.40 ±.09 95.7-104.1Values given are means + SD.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items