Effect of raw material quality on lipids and flavourcharacteristics of dried salted sardines

Author:Ariyani, Farida

Publication Date:1998

DOI:https://doi.org/10.26190/unsworks/6347

License:https://creativecommons.org/licenses/by-nc-nd/3.0/au/Link to license to see what you are allowed to do with this resource.

Downloaded from http://hdl.handle.net/1959.4/59027 in https://unsworks.unsw.edu.au on 2022-07-29

THE UNIVERSITY OF NEW SOUTH WALES

FACULTY OF LIFE SCIENCES

DEPARTMENT OF FOOD SCIENCE AND TECHNOLOGY

EFFECT OF RAW MATERIAL QUALITY ON LIPIDS AND FLAVOUR

CHARACTERISTICS OF DRIED SALTED SARDINES

A thesis submitted in fulfilment

of the requirements for the degree of Master of Science

in Food Science and Technology

by

FARIDA ARIYANI

Ir (Agricultural Technology, Gadjah Mada University)

MAppSc (Food Technology, UNSW)

Submitted

Sydney, September 1998

UNSW

1 6 JUL 1999

LIBRARY

TABLE OF CONTENTS

TABLE OF CONTENTS i

LIST OF ABBREVIATIONS vii

ACKNOWLEDGEMENT xi

ABSTRACT xii

1 INTRODUCTION 1

2 LITERATURE REVIEW 5

2.1 Fish 5

2.1.1 Fish of commercial importance 5

2.1.2 Chemical composition of fish 5

2.1.3 Fish quality 10

2.1.3.1 Factors that affect quality 10

2.1.3.2 Spoilage of fish 13

2.2 Dried Salted Fish 17

2.2.1 Processing 17

2.2.1.1 Salting 18

2.2.1.2 Drying 20

2.2.2 Effect of salting and drying on nutritional properties of fish 22

2.2.3 Quality and consumers acceptance of dried salted fish 25

2.2.4 Flavour of dried salted fish 27

2.3 Fish Lipid 28

2.3.1 Characteristics of fish lipid 28

2.3.1.1 The occurrence and distribution of lipids in fish 28

i

2.3.1.2 Lipid classes 31

2.3.1.3 Fatty acid composition 32

2.3.1.4 Flavours and odours of fish lipid 37

2.3.2 The nutritional significance of n-3 PUFA in fish lipids 37

2.3.3 Fish oil deterioration 39

2.3.3.1 Lipolysis 39

2.3.3.2 Autoxidation 42

2 3.3.3 Treatments to prevent deterioration 47

2.3.4 Methods for lipid extraction 52

2.3.4.2 Common extraction methods using single or multiple solvents 53

2.3.4.2 Chloroform-methanol solvent system 55

2.3.4.3 Hexane-isopropanol solvent system 57

2.3.5 Analytical methods to determine fish lipid stability and flavour 58

2.3.5.1 Lipid stability 58

2.3.51.1 Peroxide Value (PV) 58

2.3.5.1.2 Thiobarbituric Acid Reactive Substances (TBARS) 59

2.3.5.1.3 Polyene index (PI) 63

2.3.5.1.4 Fluorescent products 65

2.3.5.1.5 Weight gain 67

2 3.5.1.6 Rancimat method 68

2.3.5.1.7 Headspace analysis 70

2.3.5.2 Flavour 70

2 3.5.2.1 Isolation and concentration of aroma compounds 71

2.3.5.2.1.1 Static headspace analysis 71

2.3 5.2.1.2 Dynamic headspace analysis 71

ii

2.3.5.2.1.3 Distillation/extraction 74

2.3 5.2.2 Separation of aroma components 75

2.3.5.2.2.1 Column 75

2.3.5.2.2.2 Inlet system 76

2.3.5.2.2.3 Detector 77

2.3.5.2.3 Odour characterisation of individual flavour compounds 79

2.3.5.2.4 Identification of the individual volatile compounds 80

2.3.5.2.4.1 Mass spectrometry 81

2.3.5.2.4.2 Gas chromatography-Mass spectrometry (GC-MS) coupling 83

MATERIALS AND METHODS 85

3.1 Materials 85

3.2 Equipment 85

3.3 Chemicals and Solvents 86

3.4 Selection of Extraction Method for Fish Lipids 87

3.4.1 Raw materials 87

3.4.2 Methods 88

3.4.2.1 Extraction methods 88

3.4.2.1.1 Bligh and Dyer (1959) method 88

3.4.2.1.2 Hara and Radin (1978) method 88

3.4 2.2 Preparatory method 89

3.4.2.2.1 Effect of raw material preparation 89

3.4 2.2.2 Optimisation of Hara and Radin method 89

3.4.2.2.3 Evaluation of Bligh and Dyer and Hara and Radin methods 90

3.4.2.3 Analytical methods 90

3.4.2 3.1 Moisture content 90

3.4.2.3.2 Lipid content 91

3.4 2.3.3 Soluble protein 91

3.4.2 3.4 Polar compounds 91

3.4 2.3.5 Fluorescent products 92

3.4.2.3.6 Weight gain 92

3.5 Simulation of Fish Spoilage 93

3.5.1 Raw materials 93

3.5.2 Methods 93

3.5.2.1 Preparatory methods 93

3.5.2.2 Analytical methods 93

3.5.2.2.1 Sensory assessment 93

3.5.2.2.2 Moisture content 94

3.5.2.2.3 pH 94

3.5.2.2.4 Total volatile base nitrogen (TVB-N) 94

3.6 Lipid Characterisation of Raw Materials and Their Dried Salted Products 94

3.6.1 Raw materials 94

3.6.2 Methods 95

3.6.2.1 Preparatory method 95

3.6.2.2 Simulation of fish spoilage 95

3.6.2.3 Processing of dried salted fish 95

3.6.2.4 Lipid extraction 95

3.6.2.5 Analytical methods 96

3.6 2.5.1 Trimethylamine nitrogen (TMA-N) 96

3.6.2.5.2 Lipid content 96

3.6.2.5.3 Polar and non-polar compound 96

IV

3.6.2.5.4 Fluorescence 96

3.6.2.5.5 Polyene Index 96

3.6.2.5.6 Headspace/volatile compounds 97

3.7 Characterisation of Dried Salted Fish Flavour 98

3.7.1 Raw materials 98

3.7.2 Methods 98

3.7.2.1 Preparatory methods 98

3.7.2.1.1 Volatile compounds 98

3.7.2.1.2 Sample preparation for sensory evaluation 99

3.7.2.2 Analytical methods 99

37.2.2.1 Protein 99

3.7.2.2.2 Ash 99

3.7.2.2.3 Salt 100

3.7.2.2.4 Water activity (aw) 100

3.7.2.2.5 Flavour/volatile compounds 100

3.7.2.2.6 Sensory assessment 100

3.8 Statistical analysis 101

4 RESULTS AND DISCUSSION 102

4 1 Selection of Extraction Method for Fish Lipids 102

4.2 Spoilage Pattern of Australian Sardines 112

4.3 Lipid Characterisation of Raw Materials and Their Dried Salted Products 121

4.3.1 Chemical analysis 121

4.3.2 Headspace analysis 128

4.4 Characterisation of Dried Salted Fish Flavour 142

4.4.1 Chemical analysis 142

4.4.2 Headspace analysis 146

4.4.3 Sensory evaluation 159

4.4.4 Relationship between sensory assessment and objective tests data 164

5 CONCLUSIONS AND RECOMMENDATIONS 166

5.1 Conclusions 166

5.2 Recommendations 169

6 REFERENCES 170

7 APPENDICES 207

vi

LIST OF ABBREVIATIONS

AA : Arachidonic acid

AH : Antioxidant

AO AC : Association of Official Analytical Chemists

AOCS : American Oil Chemists Society

AOM : Active oxygen method

ATP : Adenosine triphosphate

BHA : Butylated hydroxyanisole

BHT : Butylated hydroxytoluene

BOPP : Biaxially oriented polypropylene

Cl : Chemical ionisation

CSW : Chilled sea water

DAG : Diacylglycerol

DHA : Docosahexaenoic acid

DHS : Dynamic headspace sampling

DMA : Dimethylamine

DMS : Dimethyl sulphide

DMDS : Dimethyl disulphide

DMPT : Dimethylpropiothetin

DMTS : Dimethyl trisulphide

DNA : Deoxyribonucleic acid

ECD : Electron capture detector

El : Electron impact ionisation

EPA : Eicosapentaenoic acid

FA : Fatty acid

FAME : Fatty acid methyl ester

FAO : Food and Agriculture Organization

FDA : Food and Drug Administration

FFA : Free fatty acid

FID : Flame ionisation detector

FPD : Flame photometric detector

GC : Gas chromatography

GLC : Gas liquid chromatography

GC-MS : Gas chromatography - Mass spectrometry

HDL : High density lipoprotein (cholesterol)

HPLC : High pressure liquid chromatography

HUFA : Highly unsaturated fatty acid

IAFMM : The International Association of Fish Meal Manufacturers

ID : Inner diameter

IR : Infrared

IU : International Unit

LDL : Low density lipoprotein (cholesterol)

LEA : Linoleic acid

LLA : Linolenic acid

LNA : Linolenic acid

MAG : Monoacylglycerol

MIB : Methylisoborneol

MS : Mass spectrometry

MUFA : Monounsaturated fatty acid

NPD : Nitrogen phosphorus detector

NMR : Nuclear magnetic resonance

NPL : Non-polar lipid

NPU : Net protein utilisation

OA : Oleic acid

PA : Palmitoleic acid

PC : Phosphatidylcholine

PE : Phosphatidylethanolamine

PER : Protein efficiency ratio

PI : Polyene index

PID : Photo-ionisation detector

PL : Phospholipid

PLOT : Porous layer open tubular

PS : Phosphatidylserine

PUFA : Polyunsaturated fatty acid

PV : Peroxide value

RH : Relative humidity

RM : Raw material

SCOT : Support coated open tubular

SDE : Steam distillation-extraction

SDS : Sodium dodecyl sulphate

SFA : Saturated fatty acid

TAG : Triacylglycerol

TBA : Thiobarbituric acid

TBARS : Thiobarbituric acid reactive substances

TBHQ : Tertiary butylhydroquinone

TCA : Trichloroacetic acid

TCD : Thermal conductivity detector

TFRU : Tasmanian Food Research Unit

TLC : Thin layer chromatography

TMA : Trimethylamine

TMA-N : Trimethylamine nitrogen

TMAO : Trimethylamine oxide

TRF : Theoretical relative response factors

TVB : Total volatile bases

TVB-N : Total volatile base nitrogen

UNSW : The University of New South Wales

VLDL : Very low density lipoprotein (cholesterol)

WCOT : Wall coated open tubular

ACKNOWLEDGEMENTS

I thank Allah, the Most Beneficent and the Most Merciful, that finally this thesis

could be finished.

I would like to express my gratitude to Prof. Ken A. Buckle for his encouragement,

guidance and his patience during all stages of my study, and to Dr John D. Craske for his

constructive criticism, suggestion and advice during my study.

I wish to thank Dr Supamo, former Head of the Research Station for Marine

Fisheries, Slipi, Jakarta, for granting me study leave. I also thank the Australian

Government for the award of an AusAID fellowship.

My gratitude is also due to all Indonesian post graduate students in the University of

New South Wales and their families for their participation during sensory evaluation

studies.

I would like also to express sincere appreciation to Mr C. Taraborrelli and Ms E.

Emmerson for their assistance, to Mr Z. Suminski for his help and friendship and all

postgraduate students of the Department Food Science & Technology for their

encouragement and friendship during my study. Thanks are also due to Ms Christine

Locke for her assistance in the internet communication between my supervisors and I as

well as in printing the thesis document from the electronic mail files.

Finally, I would like to express my sincerest thanks and appreciation to my husband

Dr Achmad Poemomo for his continuous encouragement, support and patience during my

study with its difficulties, and to my parents in Indonesia H.M. Yantodarsono for their

moral support. This thesis is dedicated for them.

ABSTRACT

The processing, quality and flavour of dried salted fish were reviewed. Fish lipid

extraction methods and characterisation were studied. The effects of fish quality on the

lipid composition and dried salted products were investigated.

The Bligh and Dyer (1959) and the Hara and Radin (1978) lipid extraction methods

were comparable, but the former produced more stable lipid.

No significant differences were detected in yield, fluorescence, saturated fatty acid

(SAFA) and monounsaturated fatty acid (MUFA) of lipids from different quality fish,

however, marked changes were observed in yield, fluorescence, polar compounds,

polyunsaturated fatty acid (PUFA) and polyene index (PI) from their dried salted products.

Twenty four aroma notes were detected from lipid extracted from raw materials and

their dried salted products, which could be classified into 5 major notes, namely green,

sweet, oxidised, pesticide and others. Their levels increased with the decrease in fish

quality.

From dried salted fish, 18 aroma notes were detected, which could be grouped into

6 major notes, namely green, sweet, oxidised, cabbage-like, pesticide and others.

Lower quality raw materials produced higher amounts of volatile compounds in

their dried salted products, and the dominant aroma was described as oxidised. Spoiled raw

materials resulted in significantly higher amount of volatile compounds exhibiting oxidised

aroma which was parallel with the levels of total volatile base nitrogen (TVB-N) and

trimethylamine nitrogen (TMA-N).

Uncooked dried salted fish prepared from fresh fish were liked most by panelists.

Regarding the aroma, however, the panelists were not able to distinguish the products

prepared from different quality fish. Although the panelists could distinguish the aroma ofxii

fried products prepared from spoiled fish, the majority could not detect if the products

were rancid.

1 INTRODUCTION

Salting and drying is the most common fish preserving method in many parts of the

world, especially in tropical countries. In Asia, dried salted fish has long been consumed as

a staple food and as an appetiser. In Southeast Asia, about 30% of landed fish is

processed as cured fish (fish preserved by reducing water content involving salting or drying

or fermentation or a combination of the three), while in Indonesia 40% of total fish

production is processed traditionally by methods such as salting, drying, boiling, smoking

and fermenting.

A combination of salting and drying in the processing of dried salted fish results in

considerable changes in the quality and character of the raw material and losses in some

nutritive compounds including lipids due to oxidation. Flavour is the most important

parameter in determining quality and is directly related to consumer preference. Fishy

odour and flavour originating from fish lipid which has undergone oxidation often hinders

the consumption of fish. Lipid in fish tissue oxidises differently from oil in its free form due

to the involvement of other biologically active compounds, thus producing specific

flavours and odours which are sometimes completely different from those of the raw

material.

In many tropical countries, such as Indonesia, fish raw materials used for the drying

and salting process are not of prime grade, but consist of those that are not sold as fresh fish

or for other industrial purposes. For oil sardines, good-quality ones are usually processed

into canned products, medium quality fish into boiled-salted products (pindang) and lower

quality raw materials into dried salted fish or fish meal. This product condition affects the

eating habits of consumers in their preference for dried salted fish, where consumers are not

familiar with salted fish made from fresh raw material. In some countries (Zambia, Pakistan

and Ghana) poor quality dried salted fish does not seem to lead to a loss of market value,

and high quality dried salted fish does not correlate to higher market value of the product.

Consumer preferences for dried salted fish are clearly different for consumers from

developed compared to less developed countries, and also from one region to other regions

in the same country. Dried salted fish, which might be unacceptable and distasteful to

Western people, will be acceptable to or even favoured by African and Asian people,

simply because their tastes are different. Most Southeast Asians are rice-eaters who prefer

to eat rice, which is relatively tasteless, with spicy dishes, such as salted and/or dried

salted fish, and canned or frozen fish cannot replace dried salted fish as a flavouring agent.

Therefore for this reason, improving the technology of processing, which would seem

preferable to Western people, might result in the decrease or even elimination of

characteristic appearance, odour and flavour of the products resulting in less acceptable

products for the Asian consumer.

The effects of processing and storage on the quality of some dried salted fish, as

well as the lipid oxidation and stability of dried salted fish, have been investigated.

However, little information on the character of fish lipid from dried salted fish prepared

from different qualities of raw material, as well as the flavour of such dried salted fish, is

available. For this reason, further study is still needed in order to acquire a better

understanding of the relationship between raw material quality and that of the

corresponding dried salted products in respect of sensory acceptance and chemical quality.

Additionally, study of the character of lipid extracted from different qualities of raw

material as well as their corresponding dried salted products will be valuable in

understanding the changes of lipid due to processing and the contribution of the lipid to the

quality of dried salted products.

Fish lipid is one of the main substances contributing to the odour and flavour of fish

2

and fish products, especially dried salted fish, and thus should be accounted for in the

characterisation of flavour and odour. In this characterisation, the lipid extraction method

should quantitatively extract all lipids from the tissue matrices in undegraded condition, as

well as give no contamination of non-lipid components so that accurate results can be

obtained.

There are many methods of lipid extraction, but the simplest and most rapid method

is preferred in extracting fish lipid, which is very susceptible to oxidation. The Bligh and

Dyer (1959) method using chloroform-methanol solvent has been considered as a rapid and

simple method and suitable to extract lipid from fish.

Another method developed in 1978 by Hara and Radin used the less toxic solvents,

hexane and isopropanol. This solvent system consists of a low-polarity solvent and a high

polarity, water-miscible solvent, which can penetrate cell membranes and dissolve a wide

range of lipids that differ markedly in their solubility. Although this method has been

reported comparable to the Bligh and Dyer method, and even more efficient in extracting

acidic phospholipids, little information is available concerning the use of this method to

extract lipid from fish.

This study deals with an evaluation of these extraction methods for fish lipid,

characterisation of fish lipid extracted from different qualities of raw material (Australian

sardine) and from dried salted fish prepared from the same raw material, as well as the

characterisation of their flavour.

This study comprised the following steps:

* selection of an extraction method suitable for fish oil,

investigation of the spoilage pattern of Australian sardine during incubation at 28-

30°C and RH 70-80%, in order to obtain degraded raw material for production of

dried salted sardine,

3

characterisation of lipid extracted from different qualities of sardine (fresh,*

moderately spoiled, spoiled) and their dried salted products, and

characterisation of the flavour of dried salted fish prepared from different raw

material qualities.

4

2 LITERATURE REVIEW

2.1 Fish

2.1.1 Fish of commercial importance

Fish constitute approximately half of the total number of known living vertebrates

and comprise over 24,000 species with an increase in the number of species defined every

year (Nelson 1994). According to Love (1982), modern fish can be classified into three

main classes, i.e. Cyclostomata, Selachii and Pisces. The Cyclostomata includes a large

number of fossil forms and only a few survive today. They are no longer of commercial

importance. The Selachii class, which are also referred to as Elasmobranchii, includes

sharks, dogfish, skates and rays. The bony fish, which are the majority of the

commercially important species such as cod, haddock, herring and halibut, belong to the

Pisces class.

Some fish of commercial importance and their usual utilisation are shown in Table

2.1, while the landings of some important species which reached more than 1 Mt are

presented in Table 2.2.

2.1.2 Chemical composition of fish

The main chemical components of fish meat are water, crude protein and lipids,

while carbohydrates, vitamins and minerals are the minor components. The nutritive value,

functional properties, sensory quality and storage stability of fish are mainly affected by the

major components, while the minor components play a significant role in biochemical

processes (Sikorski, Kolakowska and Pan 1990b).

The chemical composition of fish varies depending on species, age, sex and season

with considerable changes taking place in the water and lipid content of fatty fish

5

Table 2.1 Fish families of commercial importance

Family Habitat and nature Important species Usualutilisation

Clupeidae and Engraididae

pelagic, cold & warm waters

sprat, pilchard, anchovy, menhaden

food, feed, industrial oil

Gadidae bottom, predatory, cold waters

cod, haddock, pollock, hake

food

Serranidae tropical & warm waters gray and red grouper, marble rockfish

food

Carangidae pelagic, temperate & warm waters

horse and jack mackerel food

Lutianidae predatory, tropical waters gray & John’s snapper food

Sparidae Central Atlantic waters white stumpnose, common & red bream

food

Scienidae warm coastal waters Japanese meagre, gray weakfish, croaker, drum

food

Nototheniidae demersal, cold sub-Antarctic waters

marbled & blue notothenia

food

Scorn hr idae predatory, pelagic, warm and tropical waters

mackerel and tuna food, game fishing

Scorpenidae predatory, temperate warm waters oncontinental shelves and slopes to ~ 900m deep

ocean perch, Pacific rosefish

food

Bothidae, Plenronectidae and Sole idae

predatory, cold, temperate & warm zones in shallow & deep waters

northern fluke, turbot, halibut, sand dab, plaice, sole

food

Salmonidae predatory, migratory, spawning in fresh waters

salmon food, game fishing

Selachii pelagic, predators, warm waters

shark and spiny dogfish food

After Rutkowichz (1982)

6

(Huss 1988). During the season, when the fat content is lower, the water content increases.

Table 2.2 Total catch of fisheries by groups of species in 1993

Group of species Total catch (000 t)

Herring, sardine, anchovy 21,438

Miscellaneous marine fish 10,534

Jack, mullet, sauries 10,114

Cod, hake, haddock 9,899

Carp, other cyprinid 8,074

Redfish, bass, conger 5,648

Miscellaneous freshwater fish 5,500

Tuna, bonito, billfish 4,655

Mackerel, snoek, cutlass 3,900

Shrimp, prawn 2,893

Squid, cuttlefish, octopus 2,760

Clam, cockle, arkshell 1,816

Sea-spider, crab 1,738

Salmon, trout, smelt 1,696

Scallop, pecten 1,449

Mussel 1,192

Oyster 1,121

Flounder, halibut, sole 1,116

After FAO (1995)

In certain seasons when the water content is low, the fat content increases. The

sum of water and lipid is constant, and normally constitutes around 80% of the fillet mass

(Huss 1988, Pigott and Tucker 1990).

The water content of fish muscle varies depending on species and nutritional status

of the fish, however, generally water content in the fish is in the range of 74.8-77.2 %

(Stansby 1963).

7

There are 3 groups of fish muscle proteins, i.e. structural proteins or myofibrillar

proteins which comprise 70-80% of the total protein; sarcoplasmic protein, which accounts

for 25-30%; and connective tissue protein which amounts to 3-10% of the total protein

(Huss 1995). In the processing of meat, myofibrillar and connective tissue proteins play an

important role in coagulation and gel formation, while the tensile strength of the muscles

and the rheological properties of the meat are due to the contribution of connective tissue

proteins (Sikorski etal. 1990b, Suzuki 1981).

Fish tissue also contains small amounts of non-protein nitrogen-containing

compounds such as trimethylamine (TMA), trimethylamine oxide (TMAO), nucleotides,

free amino acids and urea (Huss 1995, Sikorski et al. 1990b). Although this fraction occurs

in small proportion (approximately 0.5 to 1.0%), it contributes to the character of fish

flavour (Jones 1967, Kawai 1996). As a degradation product of TMAO, TMA contributes

to the characteristic fishy odours, while nucleotides mainly produce a meaty taste. The

sweetness and bitterness are due to some free amino acids in fish muscle. Urea, that occurs

relatively abundantly in cartilaginous fish such as shark and ray, gives a bitter taste and

ammoniacal odour.

Lipid content of fish varies widely depending on the species, season, body part and

maturity. Lipids in most fish species consist of triacylglycerols (TAG), but in some species

such as shark, a significant quantity of the lipid consists of the hydrocarbon squalene

(Nichols et al. 1994). Fish lipids consist of fatty acids with long-chain length (up to 22

carbon) and are highly unsaturated (Ackman 1980), different from lipids from warm­

blooded animals or plants. The characteristics of fish lipid, including their distribution and

occurrence, classes and fatty acids, are discussed in Section 2.3.1.

The availability of carbohydrate in fish is limited and not significant in fish although

certain shellfish such as clams, mussels and oysters reserve their energy as glycogen. Since

8

glycogen is hydrolysed rapidly after catching, the amount of this compound in fish after

capture is very small (Love 1982, Pigott and Tucker 1990).

In general, fish is a good source of vitamins A, B and D. In seafood muscle tissue,

the vitamin B complex and vitamin C (water soluble vitamins) are available at approximately

the same amount as in animal muscle meats (Pigott and Tucker 1990). Tuna and salmon are

rich in pyridoxine, while anchovy, herring, pilchard and sardine contain a high content of

vitamin B12 (Kinsella 1988). Fat-soluble vitamins are available in relatively high

concentration in the liver of some species, such as cod and halibut, and are present in

moderate amount in the flesh of fatty fish, such as menhaden and mackerel. Since these

vitamins are soluble in fat, and the fat content of fish varies depending on the species,

maturity and season, the amount of these vitamins also varies greatly between species as

well as within species. The amount of vitamin A in the flesh of lean fish is in the range of

25-50 IU per lOOg meat, while fatty fish may contain 100-4500 IU per 100 g meat (1 IU vit

A = 1 pg carotene) (Sikorski et al. 1990b). In species such as sardine and tuna, the

content of vitamin D is in the range of 530-5400 IU (13.25-135 pg) and 700-2000 IU

(17.5-50 pg) per 100 g meat, respectively. Tocopherol is also present in moderate amounts

while vitamin K is present in small amounts in most fish (Izumi 1993, Pigott and Tucker

1990, Sikorski et al. 1990b).

Seafood is a good source of mineral components, both macroelements such as

sodium, potassium, calcium, magnesium, and phosphorus, and microelements such as

fluoride, zinc, copper and iron. Canned fish, since it is usually consumed with their softened

bones, is an excellent source of calcium and magnesium (Izumi 1993). As high- protein

food, fish is also a good source of phosphorus. Seafood is relatively rich in potassium, with

fish providing 300-400 mg/lOOg and shellfish 200-300 mg/g, but they are considered as a

low sodium food, i.e. 120 mg/lOOg for fish and 188 mg/lOOg for shellfish (Pigott and

9

Tucker 1990, Sikorski et al. 1990b). However, processed fish products usually contain a

high concentration of sodium from the ingredients, such as salt, which are added during

processing.

Iron and copper are important microelements found in fish. While red meat is

considered as the best source of dietary iron, fatty fish and molluscs are also a good source

of iron and copper, especially dark muscle, which is commonly abundant in pelagic fish.

These elements play an important role in the oxidation of fish lipid during the preparation

and processing of fish.

The nutrient composition of some common marine species in comparison to other

meat products is presented in Table 2.3.

2.1.3 Fish quality

2.1.3.1 Factors that affect quality

There are two major factors that influence fish quality, i.e. intrinsic and extrinsic

factors (Connell 1980, Pedroza-Menabrito and Regenstein 1990). The intrinsic factors

correlate more to the nature and habitat of the fish at the time of capture than to

intervention by humans.

Species, size, season, and fishing grounds are the most important intrinsic factors.

Some species of fish are preferred and are more highly valued than others. The flesh

appearance of some species, taboos and the presence of excessive numbers of bones

influence consumers' preference. Species such as halibut or salmon, as well as some

snapper having brilliant colouration, are ranked as high quality species, which makes them

expensive (Connell 1980), while small pelagic species that are commonly bony and have a

characteristic strong fishy flavour have reduced appeal to some consumers (Anon. 1991),

resulting in low price.

10

In relation to the size of fish, there is no direct correlation between the quality of

fish and the size, however, within species, the larger fish commonly provide a higher value.

This may be due to the high edible proportion of larger fish over smaller ones.

Regarding the shelf life of fish after catching, it is commonly noted that flatfish keep

longer than round fish, large fish keep longer than small fish, lean fish keep longer than fatty

fish, and bony fish keep longer than cartilaginous fish (Huss 1995). Although the reasons

for these differences remain unclear, the longer time of rigor mortis as well as the lower

post-mortem pH of large fish, such as halibut, may explain their longer shelf life (Sikorski,

Kolakowska and Burt 1990a).

Physiological changes due to the reproductive cycle certainly affect the

characteristics of most species of fish. During spawning, reserve food in the flesh and in the

liver (for some species) is transferred for the development of the gonads and, in this period,

as well as for a certain period afterward, most fish do not feed. This results in depletion of

fat, protein and carbohydrate and, as a consequence, the quality and nutritional value of fish

are reduced (Connell 1980, Pedroza-Menabrito and Regenstein 1990). When white fish of

poor flesh condition are cooked, the texture becomes soft and gelatinous and watery for

certain species. In the case of fatty fish, the high fat content is desirable for canning and

kippering in order to obtain the best quality product with an excellent appearance, and very

smooth and succulent flavour. At low level of lipids, the products become dry, fibrous or

coarse and the taste is disappointing (Love 1994).

Table 2.3 Chemical composition of some marine fish*

>. O) -Co

co n

003

CT3O

-o

3 ~

;q. 3

c0 —

o 3

cl

't

(N

CO

O

O d

d CM

0C

<D

CO CD

d d

doo cq h~

t- co eg(O

T-'

O

OJ

O

T-

oo

oo

oo

oo

oo

oo

oo

od

od

dd

dd

dd

dd

dd

dd

CM

CM

CO CO

O

CDTt

co in

t-

inq

sNcjsin

qo

o)c

ow

o)r-o

^■'oioiscoro^dr-'T

-cociiNoitN

i

CN

ICM

LO

TfC

NlT

—

T—

T— T—

T— O

'^'T

— C

OC

Ododdddddddddddd

r^c\iN

on

inio

r-qo

'_oooM

doddddddddddddd

ooooooooo^d^o^o

ddddddddd

d

d

(M'-o

iDQ

in'tn

o

1- to

T-

T-

N-

qq

qq

qq

qq

qq

qq

qo

qd

od

dd

dd

dd

dd

dd

dd

oo

T-O

Ttc

vjC

MO

JC

Din

oo

incD

CD

T-c

o't^C

OC

Oit^^O

J^CO

^MC

NC

OC

N

oo

oo

oo

oo

oo

oq

qo

qd

od

dd

dd

dd

ddd

dd

din

Tfto

iDO

istD'-o

cO'to

oo

inco

CO

t— (\|i

—

t—

t—

CN

CM

CM

CM

q q

insin

qT

f^qqoqcpN

oiD

d d

d d

d o' d

d o

O

O d

CO

qq

qq

qq

qq

qq

qq

oo

od

od

dd

dd

dd

dd

dd

dd

inN

tDO

^No

iino

ifiiDC

DN

Nn

C

MC

Mt— C

\|t— C

\|t-t-C

MC

\IC\|t—

t-

t—

t-

qq

qq

qq

qq

qq

qq

qo

q

cgdcM’d

iDC

ML

oin

dd

dd

oo

du

o

r- o

j n co Tt m

s in

t- t- ^

in

T-; Tf f- qd d d d

-b d

ddo

t-

t- co

o

d o

d d

ino^®C

NS

CT

)co9inNS

O(N

coO

(M®

"c0^^C

M^<

-OO

T-T

-r-

CO

OO

CO

C'l'tO

fflO

)C

OC

'llO't(D

CO

|.d

uD

dcr>

dc\ic

do6ddddu->

co

CTcCM

CNjT-T-CMCM

T-T-T-CgCNT-T-T-

ill°

S o

0 12

<9 ^ CO ^

E00j5ro0

"8 0

-8 ©

.rtf c$0Q

.

0O)

C/5 ~

=J CL O

•gdS'

cm cm

cd co

d d d

d

00

00

c

c: c

cz

T-; q 00 00 in in to co

q q eg

eg d d d

d

di d> d> d>

o o o

o

co

s o

o

g' eg g-

0

q q o

o

d d d

d

o

t- eg t-

co co eg

eg

o q o

o

d

uri lo

in g in in g

CO 00 CO

CM CM

o' CM d

o q q

q

d d d

d

g

cm cm

g

q q q

q

g

in in co

CO CO

T-

T—o d d

d

g

co o

cd oi oi

n g

CD 00

05 CO

cci N N

N

o. a. S o

>> 8 £

co co O

O

ll co O

co.b

c'oo&X3

C

Oc* §£ 2

szg>£c0odO

EotsoCLo2•5ID

OOucdC/><D3

JU•s03C*2caT

3IIb

c

12

1 IU vit A = 1 jag carotene = 0.6 P-c; 100 IU vit D = 2.5 pg Source: Izumi (1993)

The quality of fish is also influenced by habitat temperature as well as fishing

ground. It is generally suggested that fish from tropical waters have a longer shelf life in

ice than fish from temperate waters (Huss 1995, Sumner and Magno-Orejana 1985). In

individual species of fish or closely related species, fish caught from different fishing

grounds exhibit different qualities, e.g. hake caught from American waters exhibited a

different spoilage pattern from that caught from Argentinean waters (Huss 1988). Some

species of fish from fresh water such as rainbow trout, channel catfish and carp, have a

muddy-earthy flavour, while fish caught from polluted waters absorb the contaminants

into the body producing off-flavours such as taints due to phenolic and mineral oil

contamination (Arganosa and Flick 1992). Besides producing off-flavour, fish caught

from polluted waters often contain heavy metals such as mercury, cadmium and lead

causing toxicological risks.

The external factors causing quality loss in raw fish are mostly caused by human

carelessness during fishing, on-board handling and storing of fish. Since processing only

preserves, but does not improve the quality of raw material, the most important factor in

fish processing is to obtain good quality raw material. This can be done by proper

handling, preserving and storing of fish from the time of catching to the time of

marketing or processing.

2.1.3.2 Spoilage of fish

Fish are one of the most perishable of foods. Compared to other flesh foods, fish

contain higher amounts of protein and water, but a lower content of connective tissue,

leading to more rapid spoilage of fish.

Spoilage of fish, as in the case of other muscle foods, results from the

13

degradation of fish tissue due to enzymic and biochemical changes and microbial growth

(Connell 1980, Pedroza-Menabrito and Regenstein 1988, Ashie, Smith and Simpson

1996). As soon as fish die, a series of complex changes in fish flesh start to occur. The

process is initiated by autolysis, in which endogenous enzymes actively decompose the

flesh into simpler products. Microorganisms originally present in fish (in guts, gills and

skin) then start to multiply rapidly and putrefactive bacteria produce sulphur-containing

products having unpleasant odours and toxins (Hultin 1991).

According to Hobbs (1982), soon after catching, the fish begin to stiffen, a

process known as rigor mortis After several hours (depending on the biochemical state

of muscles at the time of death and the environmental conditions), the rigid fish gradually

softens and becomes limp again. Some of the enzymes present in the fish continue to

perform their functions. Due to the activity of intrinsic flesh enzymes, the compounds

responsible for the desirable sweetish, meaty and characteristic fish flavours are changed

to more neutral-tasting compounds resulting in a more insipid taste. When the process

of this autolysis continues, the bitter substances such as hypoxanthin (the end product of

the series of enzymatic breakdown reaction of adenosine triphosphate/ATP) may be

formed contributing to the bitter character of unfresh fish (Connell 1980, Hobbs 1982).

Hypoxanthin is commonly used as an indicator of loss of freshness (Pedroza-Menabrito

and Regenstein 1990, Colby, Enriquez-Ibarra and Flick 1993).

The viscera (gut) also contain enzymes responsible for digesting food during the

life of fish. When fish die, these enzymes attack the gut wall and surrounding tissues

resulting in bursting of the belly walls of ungutted fish, especially those that have been

feeding heavily. This commonly occurs in pelagic species (Pedroza-Menabrito and

Regenstein 1988). In some shellfish such as lobster and shrimp the digestive enzymes

attack the flesh very rapidly within a few hours after death. For that reason, most

14

shellfish should be processed quickly.

The enzymes in the muscle and the gut can break down the protein of the flesh to

smaller compounds, however, this process is quite slow in gutted fish at low temperature

(Hobbs 1982).

During autolysis, visceral lipolytic enzymes hydrolyse fish lipid resulting in an

increase in free fatty acids. Lipid hydrolysis can also be promoted by bacterial lipases

during fish spoilage. The consequences of lipolysis on fish product acceptability are not

clear, however, they may contribute to texture deterioration in combination with protein

decomposition (Sikorski et al. 1990a).

Besides lipolysis, other important changes that occur in the lipid fraction are

oxidative processes. Although autoxidation is primarily non-enzymic in nature, enzymic

degradation involving microsomal enzymes and lipoxygenase has been reported. The

degradation of lipid especially in fatty fish causes detrimental effects such as rancidity

and discoloration. The deterioration of fish lipid involving lipolysis and oxidation is

discussed in Section 2.3.3.

Although autolytic activity predominates in the early stage of fish deterioration,

the spoilage of fish is primarily due to bacterial action. The microfloral composition of

fish reflects the microflora of the area where the fish were caught. The microflora of

cold water fish is dominated by Gram-negative bacteria such as Pseudomonas,

Flavobacterium and Achromobacter, while Gram-positive bacteria such as Micrococcus

and Bacillus are dominant in warm water fish (Liston 1980).

At the time of death, the microorganisms are free to invade and diffuse into the

flesh through the intestine to the body cavity and belly walls, or through the skin into the

flesh, or through the gill tissue and proceed along the vascular system, through the

kidney and then into the flesh (Pedroza-Menabrito and Regenstein 1988). The number

15

of microorganisms in the flesh grows slowly initially, but then increases rapidly. Low

molecular weight compounds, such as carbohydrates, free amino acids, small peptides

and lactic acid, are initially used as energy sources to promote the growth of

microorganisms in fresh fish. When glucose in the fish muscle has been utilised and

depleted, utilisation by microorganisms of lactic acid and non-protein compounds,

especially TMAO, which is commonly found in large amount in marine fish, starts

(Colby et al. 1993).

The TMAO is broken down to TMA, possibly by endogenous enzymes in fish,

but mainly by enzyme activity of Pseudomonas and Alteromonas (Sikorski et al. 1990a).

When TMA reacts with lipid in the fish muscle, the characteristic fishy odour of low

quality fish is produced, thus the increase in TMA is commonly used as a chemical

measure of fish spoilage. The accumulation of ammonia and volatile fatty acids seems to

be due to oxidative deamination of amino acids. The mixture of TMA, DMA,

methylamine and ammonia, known as total volatile bases (TVB), is also used as an

alternative to measure the degree of spoilage (Connell 1980, Sikorski et al. 1990a).

The high concentration of free amino acids in the early stages of spoilage

apparently inhibits proteinases, resulting in an insignificant impact of proteolysis

activities, however, in the later stages of spoilage, proteolysis become more important

when free amino acids have been depleted (Pedroza-Menabrito and Regenstein 1988).

During proteolysis of fish proteins an increase in amino acids and volatile sulphur

compounds, such as mercaptans and hydrogen sulphide, is noted. At this stage, sensory

characteristics of spoilage are characterised by offensive odours associated with the

presence of volatile sulphur compounds.

In addition to the effect on odour and flavour, microbial spoilage also gives an

impact on the appearance and physical properties of several components of the body

16

(Connell 1980). The slime on the skin and gills that is initially watery and clear becomes

cloudy, clotted and discoloured, the skin loses its brightness and becomes dull and the

peritoneum becomes dull and easily detached from the internal body wall, therefore,

sensory attributes involving sight, touch, smell and taste are those most commonly used

as fish freshness indices (Connell 1980).

In order to assess of the degree of spoilage, scoring systems are often used such

as that developed by the Torry Research Station, Aberdeen, Scotland, in which the

score decreases for decreased quality of product, and that developed by the Tasmanian

Food Research Unit (TFRU), in which the score increases as deterioration increases

(Howgate 1982, Branch and Vail 1985).

Another important bacterial metabolite in some fish is histamine (Pedroza-

Menabrito and Regenstein 1988, Sikorski et al. 1990a). In scombroid fish such as tuna

and mackerel, certain bacteria especially Proteus morganii and Enterobacter aerogenes

can decarboxylate histidine and produce histamine that, at certain levels, can be toxic for

humans. The optimum temperature for histamine accumulation in fishery products

generally is in the range of 20-45°C, however, it also depends on the presence of

histamine-forming bacterial species and the properties of the fish. The formation of

histamine at refrigeration temperature is usually negligible (Reilly and Santos 1985).

2.2 Dried Salted Fish

2.2.1 Processing

Salting and drying is the most common fish preservation method in many parts of

the world, mainly in tropical countries to suit the prevailing climate (Howgate and

Ahmed 1972, Bostock, Walker and Wood 1987). In Asia, dried salted fish is a staple

food and appetiser (FAO 1981, Maynard 1983, Manurung 1995, Yu 1995).

In tropical countries, high ambient temperatures, coupled with a lack of ice, result

17

in very rapid spoilage of fish. Fish which cannot be sold fresh are usually preserved by

traditional curing methods, such as salting and drying (Yu 1995). The preservation is

basically achieved by reducing the water activity (aw) to a level at which spoilage is

retarded or at least greatly slowed. This is achieved by adding salt to render the moisture

unavailable to microorganisms and/or by drying the fish to remove water directly from

fish.

2.2.1.1 Salting

The aim of salting is to extract the water and increase the concentration of

solutes to minimise the amount of water available to spoilage microbia in the flesh.

Chloride also gives some detrimental effects on microorganisms.

Salting methods can be divided into 3 types, i.e. dry or ‘kench’ curing salting,

wet methods, i.e. brining, and a combination of dry and wet methods which is called

‘pickle’ salting (FAO 1981, Horner 1992). In dry salting, fish are arranged in alternate

layers with salt crystals and the brine formed is allowed to drain. This method is

commonly applied for white and lean fish with a total lipid content not more than 1%. If

the brine formed in dry salting is not allowed to drain so that the fish become immersed

in a strong brine, or when fish is salted simultaneously with salt and brine, the method is

called ‘pickle’ salting. Fish with a high lipid content, including mackerel, sardine and

anchovy, are recommended to be salted by a combination between dry and wet methods.

By immersing the fish in brine, oxygen entrance is restricted, resulting in the retardation

of rancidity.

In wet salting, fish is immersed in salt solution, which is usually saturated brine.

This method is commonly applied for flavouring purposes or to produce an attractive

surface in smoked fish products, by which fish is salted for a short time in a less

saturated brine before smoking (Horner 1992).

18

Since salting is a race between the action of salt and deterioration, the size and

the form of fish determines whether fish are salted whole and ungutted, gutted and

butterfly-shaped, or in smaller pieces ranging from fillets to mince. Small species such

as anchovy, small herring and small sardine can be salted whole without gutting,

however, larger fish salted in this manner deteriorate at the centre before salt takes

effect. Therefore, larger fish must be salted in cut forms and medium size fish such as

grouper, sea bream and croaker should be at least gutted and salted in halves or

butterfly-shaped.

The purity of salt is very important since it affects the physical character and

colour of the fish products. Calcium and magnesium salts, although they produce whiter

products, tend to impart a bitter flavour. Besides, their hygroscopicity is also high which

causes the end product to be surface damp or wet (Burgess et al. 1965, Wheaton and

Lawson 1986). In addition, undesirable brown or yellow colour may occur if heavy

metals such as copper and iron are present; although at only low levels (a few mg/kg)

they also accelerate the development of rancidity in fish that contain a high fat/oil

content (van Klaveren and Legendre 1965, Wheaton and Lawson 1986).

The amount of salt to be added varies depending on the purpose of the process

and the geographical area. However, it is recommended that in dry salting the

proportion of salt to fish is 3:10 (w/w) (Burgess et al. 1965), although for many

practical operations a greater quantity of salt can be used ( van Veen 1953, Nasran et al.,

1992). In wet salting, at least 1:1 (w/w) ratio of brine to fish is recommended when

using saturated brine, and it should be ensured that a certain amount of solid salt is

always present in the bottom of the brining barrel (Burgess et al. 1965).

The time of salting also varies and is clearly affected by the thickness of fish,

brine concentration and fat content (Jason 1965, Berhimpon et al. 1990, 1991). In

19

practical operations carried out by traditional processors, salting is commonly carried out

overnight either for dry or wet salting, but the salting time could be longer whenever the

weather is unsuitable for sun drying .

2.2.1.2 Drying

Salting alone is not sufficient to produce long shelf life products, therefore

subsequent drying of salted fish is necessary. The principle of drying in this case is

removal of water from the product by evaporation, although other methods such as

applying pressure or adding absorbents are also known as drying.

Removal of water from fish during drying consists of 2 stages. The first stage,

termed the constant-rate period, is evaporation of water on the surface of fish. When

surface water has evaporated, the surface starts to dry and diffusion of the water within

the fish to the surface of the fish takes place, which starts the falling-rate period (Burgess

et al. 1965, Jason 1965). During the first stage, the rate of drying is constant and is a

function of temperature, velocity and relative humidity of air, as well as the surface area

of the product. Higher air temperature and velocity and lower relative humidity, result in

faster water evaporation (Doe and Olley 1990). The drying rate is also faster when the

fish is thinner and leaner (Waterman 1976).

The second falling rate stage is not much affected by relative humidity and

velocity of air, but is directly affected by the internal factors of the product itself, such as

the size of fish, and the fat and salt contents (Doe and Olley 1990, Waterman 1976).

The water movement from inside the flesh to the surface becomes faster in thinner and

leaner fish. The presence of a greater amount of salt results in slower water diffusion. In

this period, the rate of water evaporation is not constant.

Fish drying can be carried out either by the simple traditional method, i.e. sun

20

drying, or by improved drying processes, such as solar drying, mechanical drying and

freeze drying (Burgess et al. 1965, Doe and Olley 1990). Sun drying is the simplest and

oldest method, where fish is directly exposed to sunlight. In tropical countries, fish or

salted fish are simply placed on racks consisting of bamboo trays about 0.5 to 0.75 m

above the ground, or hung on suspended ropes. Sometimes the fish is just spread out

directly on the ground (Waterman 1976). The sun drying method, commonly noted as

natural drying, is inexpensive because energy used for drying is free. However, this

method is highly dependent on the weather. During good weather or bright days, drying

is usually carried out for less than a day for small fish. When the weather is not good

for drying, a plastic sheet is placed on the top of the racks and removed to continue

drying whenever the weather has improved Problems encountered in sun drying include

limited drying time per day (6-7h), product exposure to contaminating agents, such as

dust, sand and dirt, and liability to insect infestation, especially flies.

Artificial drying has been developed partially to solve the problems encountered

in sun drying. Some artificial drying methods include solar drying, agrowaste drying and

mechanical drying, which have been introduced and tested in some countries (Doe and

Olley 1990). Solar drying uses the sun as the energy source, which is collected or

concentrated upon glass or plastic sheeting and an air flow is induced through natural

convection. The dried fish produced with this method is often better in quality compared

to that from natural drying (Doe and Olley 1990, Bostock et al. 1987). The materials

used as fuel in agrowaste drying include rice husk, coconut shell, peanut husk and

sawdust (Sison et al. 1983, Wibowo et al. 1993, Souness and Wibowo 1994), however,

with this method, difficulties in maintaining high air flow while keeping the temperature

sufficiently low to dry fish have been encountered (Doe and Olley 1990).

The mechanical drying method is the most widely used for fish drying in many

21

parts of the world Electrical heating and wood fires coupled with an electrical fan have

been used to control air temperature and humidity. Heated air is drawn over the fish,

which are hung on racks or placed on trays and either some parts or all of the exhausted

air can be recycled in this method. Since the temperature and humidity are easier to

control, the product dried with this method has a good quality (Doe and Olley 1990).

Temperature for mechanical drying varies depending on the species and size of

fish and whether the fish has been salted or not. In some tropical areas, the temperature

of the dryer has been set to 30-40°C, with humidity in the range of 50-55% and air

speed 1-2 m/sec (Waterman 1976). In Cambodia, the temperature of 43°C, relative

humidity of 50-55% and air velocity of 1.8m/sec were selected to dry fish, while in India,

sardine and mackerel were tunnel dried at 45°C with relative humidity of 50%.

Poernomo (1986), working with Australian sardine, recommended air temperature of

45°C, relative humidity of 30% and air velocity of 2m/sec to dry salted sardines.

Other methods of fish drying include freeze, rotary, spray, and vacuum drying

(Burgess et al. 1965, Jason 1965). Besides being an expensive process, fish dried by

freeze drying is worse in quality than frozen fish (Burgess et al. 1965), therefore this

method is only designed for special products such as fish soup (Doe and Olley 1990).

Rotary, spray and vacuum drying are not commonly used for whole, dressed, or filleted

fish, but are mostly used for fish protein concentrate.

2.2.2 Effect of salting and drying on nutritional properties of fish

During salting and drying, some chemical reactions occur resulting in changes in

nutritional value, such as those of protein and lipid. Heat treatment causes denaturation

of protein, the effects of which vary, depending upon the type of protein in the same

fish, or the same protein in different fish In general, at temperatures of 60-65°C, about

22

90% of protein is denatured, and denaturation may result in protein aggregation,

formation of disulphide linkages and reaction of the e-NH2 group of lysine and reducing

non-protein compounds such as sugar and lipid oxidation products (Opstvedt 1988).

Salting also causes denaturation of fish protein. Although denaturation by salt

results in effects similar to that by heat, i.e. decreased extractability of fish muscle, the

reason for denaturation itself is different whereby the Na‘ and Cf ions disturb the

negatively and positively charged groups of protein, resulting in changes in the native

conformation of protein (Sikorski and Ruiter 1994). Sikorski and Ruiter (1994) stated

further that Na4 from brine may replace Car and Mg' from muscle resulting in soft

texture, conversely, Ca4’ and Mg4 from impure salt may accumulate in the flesh

resulting in tough texture. Salting also affects the activity of proteinases differently

depending on their character, and increasing time of salting/ripening may enhance the

hydrolysis/autolysis of protein and, due to other complex reactions involving

carbohydrate and lipid during ripening, significant changes in muscle texture and flavour

may occur (van Klaveren and Legendre 1965, Shenderyuk and Bykowski 1990, Sikorski

and Ruiter 1994).

Aitken et al. (1967) working on cod salting found that light salting resulted in

product of higher net protein utilisation (NPU) than that derived from heavy salting,

while drying of salted cod at high temperature, i.e. 115°C, decreased the NPU compared

with drying at 100°C. Astawan et al. (1994a) studied the effect of high salt content of

Indonesian dried-salted tuna on rats and concluded that high salt content of dried salted

fish had no measurable effect on its protein quality based on a rat bioassay. Drying of a

lean marine fish Nempiterus japonicus at 60°C and 70°C decreased the level of

sulphydryl groups significantly, while drying at 50°C did not, and from the protein

efficiency ratio (PER) and NPU, drying at 60°C was superior to that at 70°C (Raghunath

23

eta/. 1995).

Salting and drying also affect fish lipid. The addition of salt affects the aw of fish,

due to the ability of salt solution to lower the partial pressure of water vapour. At lower

aw, lipid oxidation is accelerated (Doe and Olley 1990) Castell, MacLean and Moore

(1965) stated that the addition of 1% NaCl to blended cod promoted rancidity, and the

rate of development increased proportionally to the concentration of salt added when the

addition of salt was between 1-12%, but the addition of more than 12% salt did not

change the rate of development. In fatty fish, Nambudiry (1980) found that salt at low

concentration (8%) acted as a prooxidant, but in higher concentration it inhibited lipid

oxidation. The addition of NaCl in pure lipid, however, did not accelerate oxidation, and

the salt accelerated lipid oxidation only when it was in combination with muscle protein

break down (Castell et al. 1965). According to Shewan (1955), product browning and

rancidity is the result of trace amounts of heavy metals in salt such as Cu and Fe. Traces

of Cu in the curing salts were able to cause marked browning during salting of white

fish, while iron in salt produced some discoloration only when present in higher level

(concentration above 100 mg/kg).

Both natural and artificial drying involving relatively severe heat cause oxidation

of fish lipid. The carbonyl compounds such as ketones and aldehydes as oxidation

products can react with the e-NH2 of lysine (Maillard reaction). This Maillard reaction,

coupled with oxypolymerisation of unsaturated fish oil, results in the darkening,

toughening and bitter flavour of dried fish upon storage (Bligh, Shaw and Woyewoda

1988).

Salting and sun drying of Indian mackerel lowered the iodine value and increased

the peroxide value (PV) and TBA values indicating that lipid oxidation occurred during

the process (Rao and Bandyopadhyay 1983), while processing and storage of sun dried

24

or salted-dried Indonesian mackerel resulted in an increase of TBA and fluorescence and

significant loss of polyunsaturated fatty acid (PUFA) (Maruf et al. 1990). Smith, Hole

and Hanson (1990) reported that the onset of lipid oxidation occurred during salting and

drying of Indonesian marine catfish, which was shown by increasing fluorescence and

the level of acetic acid soluble substances. The loss of 30% PUFA during salting was

also reported. Exposure to sunlight during traditional drying of fish is believed to

contribute to the extent of lipid oxidation of dried salted fish. Davis et al. (1993) noted

an increase of initial oxygen uptake in a model system consisting of a highly

polyunsaturated fish oil with increase of light and temperature. Considerable lipid

oxidation, measured by PV and TBA, was observed during heating and drying of cod

(Gadus morrhua) and hilsa {Hilsa i/isha) in a mechanical dryer at 30°C (Howgate and

Ahmed 1972). Lubis and Buckle (1990) reported that oxidation and rancidity occurred

during storage of sun dried-salted Indonesian sardine and artificial dried-salted

Australian sardine evaluated by sensory judgment, polyene index and fluorescence.

Processing of Ghanaian fermented fish products involving salting and drying increased

PV and TBA during processing which further decreased during subsequent storage,

while PUFA decreased, accompanied by an increase in the proportion of saturated fatty

acids during the process (Yankah, Ohshima and Koizumi 1993).

2.2.3 Quality and consumers acceptance of dried salted fish

The quality of dried salted fish is clearly affected by many factors, such as quality

of raw material, quality of salt, and the handling and processing of the products.

However, judging the quality of dried salted fish is very subjective, especially in less

developed countries. It is commonly accepted that a good quality raw material will give

a good final product. However, regarding dried salted fish, this is not always true since

25

it is a completely different product with its own particular flavour and texture.

In tropical countries, fish raw materials used for salting and drying are not the

prime grade, but those that are found impossible to sell as fresh fish or for other

industrial purposes (FAO 1981, Connell 1980, Waterman 1976). In Indonesia, good-

quality oil sardines are usually canned, medium quality are processed into boiled-salted

products (pindang) and lower quality fish into dried salted fish or fish meal

(Burhanuddin et al. 1984, Putro 1986, Sugiyono 1994, Supamo et al. 1989). This

condition affects the eating habits of consumers in the preference of dried salted fish,

where consumers are not familiar with salted fish made from fresh raw material. This

phenomenon was supported by evidence that, in some countries (Zambia, Pakistan and

Ghana), poor quality dried salted fish did not seem to lose market value, and high

quality dried salted fish did not correlate to higher market value of the product (FAO

1981).

Consumer preferences for dried salted fish are clearly different, not only from

country to country, but also from one region to another region in the same country.

Dried salted fish which might be unacceptable and distasteful to Western people, can be

acceptable or even favoured by African and Asian people (FAO 1981). Within

southeast Asian countries, the most favoured product in Thailand is dried brined

gourami, while in Malaysia and Singapore dried salted threadfin fillets are favoured. In

Indonesia, the most preferred product in Java is boiled-salted fish, while in Sulawesi

(North Central Indonesia), smoked-salted fish is preferred (Maynard 1983).

The unacceptability of dried salted fish by Western people is not necessarily due

to the lower standard of acceptance of Asian and African peoples, but simply because

their tastes are different. For that reason, improving technology of processing, which is

better from a Western viewpoint, might decrease or even eliminate the characteristics

26

(appearance, odour and flavour) of the products resulting in less acceptance by the

consumers in Asia and Africa.

2.2.4 Flavour of dried salted fish

Flavour is important in determining quality and is directly related to consumer

preference. Fishy odour and flavour that originate from oxidised fish lipid (Stansby

1971) often hinders the consumption of fish. Lipid in fish tissue oxidises differently from

oil in the free form due to the involvement of other biologically active compounds

(Khayat and Schwall 1983, Ke et al. 1977, Logani and Davies 1979, Stansby 1971).

Processing of fish such as salting, pickling and drying, may cause oxidation of the lipid

in the tissue matrices producing specific flavours/odours with their own characters

(Castell et al. 1965, van Veen 1953, Soliman et al. 1983, Triqui and Reineccius

1995a/b).

Salting was reported to enhance the flavour of shellfish and significantly

increased the concentration of furan derivatives and some aldehydes (Soliman et al.

1983). Castell et al. (1965) reported that “rotten seaweed smell” is the characteristic of

dried salted cod. Cured anchovy processed by salting and maturing had a strong

characteristic flavour and two important groups of volatile compounds were determined

(Triqui and Reineccius 1995a,b) They were characterised as enzymically generated C8

alcohols and ketones along with trans,cis-2,6-nonadienal, which contributed plant- and

cucumber-like aromas, and autoxidatively derived C7 to C10 conjugated aldehydes which

imparted fatty and fried fat-like aromas.

Pedah Siam, an Indonesian traditional fish product made from Scomber sp. salted

for 12-24 h, partially dried and ripened, has a peculiar flavour contributed by methyl

ketones, butyl aldehyde and unsaturated aldehydes (van Veen 1953).

27

2.3 Fish Lipids

2.3.1 Characteristics of fish lipids

2.3.1.1 The occurrence and distribution of lipids in fish

Different species of fish have different amounts of total lipid, as well as a different

lipid distribution. Stansby (1982) classified fish into category A if they have a lipid

content less than 5%, category B with lipid content in the range of 5-15% and category

C if they contained more than 15% lipid. Bottom-feeding fish (e.g. cod), which possess

little or no depot fats, and a number of species that have only a modest quantity of depot

fats, are classified under category A, while category B includes most species that contain

fairly large quantities of depot fats such as herring, mackerel and sardine. Since many of

these latter species vary widely in lipid content from season to season, they may at one

season contain so little lipid that they are categorised temporarily as category A, or in

another season have a high lipid content and move into category C. Other classifications

divide the groups into lean fish with less than 2% lipid (such as cod, haddock and

shellfish), low lipid of 2-4% (such as halibut and flounder), medium lipid of 4-8% (such

as salmon) and high lipid of more than 8% (such as herring and mackerel) (Ackman

1989).

The distribution of lipid in an individual fish is not uniform. A thin layer

immediately below the skin is loaded heavily with fat and is called subcutaneous fat

(Burgess et al. 1965).

Lipids are found in both the light and dark muscles. Commonly the dark muscles

contain more lipid than the white muscles, often as much as twice the lipid content

(Mannan, Fraser and Dyer 1961 a,b, Ackman 1980, Love 1982). The dark muscle is

predominantly a “v” shaped strip running along the side of the fish immediately under the

28

skin. The typical cross sectional diagram of a fish body is shown in Figure 2.1. In

certain species, such as tuna, this block of dark muscle lies near the backbone.

Generally, pelagic fish, being the most active fish that swim continually in surface

waters, have more dark muscles and are fattier, while those that are less mobile contain

very little dark muscle, and flatfish, which spend much of their time on the sea bed, do

not have dark muscle (Sikorski, et al. 1990a).

The liver is also the fish organ that stores more lipid than flesh. The amount of

lipid in the flesh from some species of Papua New Guinean fish was in the range of 0.1-

5.7%, while lipid in the liver was higher, ranging from 2.6-20.6% (Hansel et al. 1993).

FL

Figure 2.1 Cross-section diagram of fish (Ackman 1980) FL: lateral dark muscle, BF: belly flap

29

Fish such as cod and haddock with low lipid content (<5%), have little

triacylglycerol (TAG) in their flesh and most of the oil in the flesh occurs as

phospholipids (PL) which are associated with the cell (Stansby 1982). The liver of this

kind of fish is usually large (10% of body weight), contains a high amount of lipid (50-

80% w/w) and consists largely of TAG. The lipid of the medium and high lipid fish

species mostly occur as TAG and is stored as globules in their muscles. The livers of this

species are small (1-3% of body weight), reddish or mottled and relatively non-oily

(<5%) (Love 1982, Sikorski et al. 1990b, Stansby 1982).

The belly flaps are also the most fatty region of many fish bodies, such as

mackerel and halibut. The lipid content of the belly flap of mackerel was found to be

about 37.2% while in the thick edible tissue the lipid content was 18.8% (Mannan et al.

1961a). In the Atlantic halibut the lipid content of the belly flap was higher (4.8%) than

in the edible tissue of the thick section (3.1%) and the tail section (1.2%) (Mannan et al.

1961b).

Generally, less than 1% of fish lipid occurs as PL associated with the cellular

structure of all tissues, while the majority of fat, which is in the form of TAG, may be

dispersed throughout the flesh or be restricted to depots. However, some fish may

contain only a very small amount of TAG sometimes under 1%. Fish, such as cod and

haddock, with 5% or less total oil, have very little TAG in their flesh, most of the lipid

occurring as PL, whereas fish containing 10% or more oil, such as herring and salmon,

have a large proportion of their lipids as TAG (Stansby 1982)

The lipid content depends on many factors, such as metabolism rate, maturity,

environmental temperature, and food availability; and the range in fatty fish may be very

large, from 1% to 25% wet weight. Therefore, the lipid content of fatty fish during

certain periods may be equal to that of lean fish, while lipid content and composition in

30

lean fish vary only slightly, with more pronounced variation being observed in the liver.

In activities such as fast movement and reproduction as well as during starvation,

the total food intake is less than that required for energy expenditure and growth, and

under these conditions, some of the tissue lipids are used to satisfy the energy

requirements (Sikorski 1990, Burgess et al. 1965). During gonad maturation, lipids are

transported from the liver and muscles to the gonads, and after spawning, the lipid

content in the gonads decreases markedly. Since the fish starts intensive feeding after

spawning, the lipid content in the liver and meat then increase.

2.3.1.2 Lipid classes

Similar to other animal and vegetable oils and fats, fish oils are a mixture of TAG

of various long chain fatty acids with small amounts of monoacylglycerols (MAG) and

diacylglycerols (DAG). In addition, various PL are also present, usually in amounts

considerably less than that of the TAG. Other classes of lipids, which are also present

in a few species, are sterols, wax esters and hydrocarbons. The major type of PL is

phosphatidylcholine (PC) occurring to the greatest extent with phosphatidylethanolamine

(PE) as the second most common class. They account for about 75% of the total PL

content in fish and the remainder is composed of phosphatidylserine (PS), sphingomyelin

and other PL classes (Sikorski et al. 1990b, Stansby 1990a).

Sterols are another important component of fish oils consisting almost exclusively

of cholesterol. The sterols in finfish are nearly 100% cholesterol, while in shellfish, the

cholesterol ranges from 40% in most mollusks to nearly 100% in lobsters and most crabs

(Pigott and Tucker 1990). Most finfish contain about 20 to 40 mg cholesterol per 100 g

meat, while shrimp contains up to 200 mg cholesterol per 100 g shrimp (Stansby 1982).

Wax esters occur in a wide range of species, mostly species not used as food.

31

Since wax esters serve to increase buoyancy in marine organisms, they accumulate in the

species that have to endure a long-lasting starvation period and those living in deep

waters (Sikorski et al. 1990b). Orange roughy and Dory oils are composed almost

exclusively of wax esters reaching levels of 97% and 90%, respectively. These

compounds occur mostly in the head, swim bladder, frame and skin (Nichols, Nichols

and Volkman 1993, Nichols et al. 1994).

Hydrocarbons are the principal component of the liver oil of sharks. The

proportion of shark liver is up to about 20% of the body weight and contains

considerable quantities of oil which is often rich in the C30 unsaturated isoprenoid

hydrocarbon squalene. In some commercially produced Australian oils (e.g. shark liver

oil), squalene accounts for up to 80% of lipid (Nichols et al. 1994).

2.3.1.3 Fatty acid composition

The PUFA are the most characteristic feature of marine lipids although they are

also found in some vegetable oils such as sunflower, soybean, linseed and canola

(Kinsella 1990). The position of the double bonds in the fatty acid chains distinguishes

the family of PUFA. For nutritional purposes, the different families of PUFA are

indicated by the location of the first double bond counted from the terminal methyl group

(n). The important families of PUFA are n-9, n-6 and n-3. Within a family, the acids

are biosynthetically related, being interconverted by enzymic processes of desaturation

and chain elongation, but no changes occur between families (Lobb 1992). Figure 2.2

shows the biosynthesis of PUFA.

Humans are not able to synthesise n-6 nor n-3 fatty acids, but can synthesise

saturated and n-9 fatty acids. The n-6 family fatty acids are made by terrestrial plants,

32

18:3 n-3-------- 18:4 n-3====20:4 n-3--------- 20:5 n-3======22:5 n-3--------- 22:6 n-3a-linolenic eicosapentaenoic acid docosahexaenoic acid

(EPA) (DHA)

18:2 n-6--------18:3 n-6====20:3 n-6---------20:4 n-6=====:=:22:4 n-6----------22:5 n-6

linoleic y linolenic di-homo arachidonicy linolenic

18:1 n-9--------18:2 n-9====20:2 n-9---------20:3 n-9======22:3 n-9oleic

Figure 2.2 Elongation and desaturation pathways in the biosynthesis of PUFAs,the sequential desaturation and elongation reactions with the 3 different families of PUFAs are designated with lines for the desaturation reactions catalysed by the delta 6, 5 and 4 desaturase enzymes and double lines for the elongation reactions (Turley and Strain 1993)

and marine plants such as algae and phytoplankton are able to make n-3 family fatty

acids. Thus n-6 family fatty acids in humans originate from plant oils and n-3 family

fatty acids are primarily from fish consuming the marine plants that make n-3 family

fatty acids (Dyerberg 1986, Pigott and Tucker 1990).

Compared to other edible vegetable oils and animal fats, fish oils have relatively

higher proportions of PUFAs with five or six double bonds, especially 20 or 22 carbon

chain length, reaching up to one-third or even one-half of the total fatty acid content.

Another unique characteristic of the fatty acids in fish oil is that the majority of PUFA

belong to the family n-3, while the major fatty acids in vegetable oils are in the n-6

position (Bimbo 1990, Pigott and Tucker 1990, Stansby 1982).

The major PUFAs in marine oils are C2o:5 and C22.6 of the n-3 family reaching up

to 15-30%, while Cig:2n-6 (linoleic acid/LEA) and Ci8:3n-3 (linolenic acid/LNA) commonly

found in vegetable oils occur at no more than 1 or 2% in fish oil (Ackman 1980, Stansby

33

1982). Minor fatty acids of polyunsaturated types include a series with n values of 1, 4,

or 7 occurring to the extent of 1-2% total fatty acid in most fish oils (Ackman 1980).

The monounsaturated fatty acids (MUFAs) are composed mainly of Ci6.i n-9

(palmitoleic acid/PA), Cig:i n-9 (oleic acid/OA) and its isomer Ci8:i n-i. These are often

accompanied by large amounts of C2o:i n-9 and C22:i n-9- Palmitic acid (Ci6:o) is the

principal saturated fatty acid in marine oils, whereas Cig:o, C2o:o, C22:o and C24:o are

present only in small amounts (Ackman 1980, Padley, Gunstone and Harwood 1986).

Fish oils also contain more fatty acids with odd carbon number such as Ci5, Cn and Ci9

than other oils and fats ranging from 1-3% (Stansby 1982). The fatty acid composition

of some fish oils compared to that of other oils and fats is shown in Table 2.4.

Table 2.4 Fatty acid composition (% w/w) of fish oils compared to

several vegetable oils and animal fats

Fatty Fats and oils originated fromacid Butter fat Com Lard Tallow Soybean Herring Menhaden Anchovy Tuna

C4 3.0 - - - - - - -C6 1.0 - - - - - -

C8 1.5 - - - - - -CIO 3.0 - - - - - -C12 3.5 - - - - - -C14 12.0 - 1.5 3.5 7.0 9.0 7.0 4.5C16 28.0 12.0 25.0 25.5 11.0 12.0 18.0 15.0 21.0

C16:1 3.0 - 3.0 4.0 - 6.0 9.0 9.0 6.0C18 13.0 2.0 13.0 19.5 4.0 1.0 3.0 3.0 5.5

Cl 8:1 28.5 25.0 45.5 41.5 21.0 10.0 10.5 10.0 16.5Cl 8:2 1.0 60.0 10.5 2.5 55.5 2.0 1.0 1.0 1.0Cl 8:3 - 0.5 1.0 8.5 1.0 1.5 0.5 1.0C20:1 - - 13.0 1.0 2.0 2.0C20:5 - - - - 9.0 14.5 18.0 6.0C22:1 - - - - 19.0 - 1.0 1.0C22:6 - - - - 8.0 12.0 16.0 23.0After Bimbo (1990a).

Since the depth, light and temperature of the seas both directly and indirectly

influence the feeding habit of fish (Sikorski and Karnicki 1990), geographic location,

food availability and season will also influence the fat content and fatty acid composition

34

of fish. Tropical fish are richer in n-6 fatty acids than Northern Hemisphere fish. The n-

3 PUFA content of Australian fish and other marine species is generally lower than that

of species from the Northern Hemisphere due to the low fat content (Brown, Roberts

and Truswell 1989, Sinclair et a/. 1992).

Compared to marine organisms, lipids from freshwater fish are richer in Ci6 and

Ci8 fatty acids, but lower in their contents of C2o and C22 fatty acids, while n-3 fatty

acids in fresh water species are lower than that in marine species. Different feeding habits

of freshwater and marine fish result in different proportions of fatty acids including n-3

PUFA . Most marine fish feed on flora and fauna from the sea and freshwater fish flesh

is naturally high in n-3. In freshwater species, the proportion of Cis:3 n-3 in the diet is

higher and leads to storage of this acid in depot fat without further chain extension,

resulting in lower proportions of long chain fatty acids such as C2o:5 n-3, C22;s n-3 and

C22:6 n-3-

Generally, a decrease in ambient temperature is accompanied by an increase in

the degree of unsaturation of fatty acids and suggested that temperature has an indirect

influence on the fatty acids via the food chain, where low water temperature will

correspond to a period of scarcity for most species especially fresh water fish (Ackman

1967). Other workers have reported that, in plankton, which is important in the diet of

marine fish, the degree of unsaturation of its fatty acids increases, but the chain length of

the fatty acids decreases with a reduction in environmental temperature (Lewis 1962).

Lee et a/. (1986) stated that the content of saturated and monoenoic acids of Korean

sardine oil tended to decrease, starting in winter (Dec-Feb), showing a minimum in April

and increasing again thereafter Polyenoic acids, such as C20:5 and C22:6 acids, showed a

maximum in April. Tables 2.5 and 2.6 present the variation of fatty acid composition of

fish oil from different geographic locations and different feeding habits.

35

Table 2.5 Fatty acid composition (% w/w) of fish oil from different areas

Fatty acidMenhaden oils Anchovy oils

Gulf of Mexico Atlantic Mexican ChileanWest East S.Vicente T.huana

Saturated 39.1 38.3 34.9 34.7 32.1 30.5Monoenes 27.1 26.4 30.4 28.5 30.2 32.6Dienes 3.3 4.1 3.9 1.4 2.7 2.8Trienes 3.8 3.8 2.2 2.0 2.5 2.5Tetraenes 5.3 5.7 5.7 4.9 7.2 6.0Pentaenes 15.7 15.8 14.0 15.5 21.0 16.9Hexaenes 5.1 6.0 7.8 12.4 4.3 8.9

After Ackman (1980)

Table 2.6 Effect of diet on important fatty acids (% w/w) in three species of freshwater fish

Fatty acid Channel catfish American eel Atlantic salmonculture wild culture wild culture wild

04:0 1.3 3.0 5.6 3.3 3.8 1.506:0 10.5 20.8 20.2 15.5 12.7 14.208:0 5.4 4.3 2.9 4.5 2.5 5.3

06:1 4.5 7.4 15.7 10.5 7.6 5.508:1 12.6 22.5 35.3 27.3 15.9 12.6C20:1 1.3 1.1 1.9 2.2 10.9 0.8C22:1 tr nd - - 8.8 0.2

08:2 n-6 0.5 0.3 0.7 8.9 8.1 3.1C20:4 n-6 5.7 2.3 1.2 4.4 0.5 8.0C22:4 n-6 1.2 3.7 0.8 2.3 0.2 1.3C22:5 n-6 1.8 nd 0.8 0.8 0.4 2.0

08:3 n-3 1.2 0.5 0.6 2.6 0.7 2.208:4 n-3 0.4 4.0 - - - -

C20:4 n-3 0.9 0.2 - - - -

C20:5 n-3 15.5 0.7 2.3 2.9 2.5 4.6C22:5 n-3 3.7 3.7 2.0 2.7 1.0 3.3C22.6 n-3 23.5 4.7 7.0 5.4 9.7 15.4After Ackman (1989)tr = trace nd = not detected

= not measured

36

2.3.1.4 Flavours and odours of fish lipids

As in the case of fish and fish products, fish oil has a characteristic flavour and

odour. The oil extracted from different species of fish has a flavour which is commonly

quite characteristic of the species. Karahadian and Lindsay (1989a) found that green

flavour, burnt flavour or cod liver oil-like flavour, and trout-like odour contributed to

the flavour of deodorised fish oils.

Reactions among fish components, such as amino acids, lipids and sugars during

extraction employing heat, may also contribute to the fishy odour of crude oils

(Fujimoto 1989). According to Stansby (1990b), fishy flavour/odour arises from a

combination of many compounds present in both fresh fish and slightly deteriorated fish

and can be classified based on the sources and their characteristics. The characteristics

of fish oil are described as burnt (B), green (G), natural species giving characteristic

flavour and odour (N), pure oxidation types occurring at an early stage of oxidation

(O), and pure oxidation types occurring at later stages of oxidation (R) referred to as

rancid (Stansby 1990b).

2.3.2 The nutritional significance of n-3 PUFA in fish lipids

Although other active factors may also be involved, the beneficial effects of fish

oils are generally considered to be related to the n-3 fatty acids, especially EPA and

DHA. The edible plant oils and the PUFAs found in animal muscle fat are primarily LEA

(18:2 n-6) and only a small amount of LNA acid (18:3 n-3). PUFAs with 18 carbon

atoms can be elongated and desaturated in all mammals including humans, but LEA and

LNA will compete with each other for the same enzymes. The n-6 family is favoured in

this competition, forming arachidonic acid/AA (20:4 n-6) (Dyerberg 1986). Thus the

conversion of LEA (18:2 n-6) to AA (20:4 n-6) minimises the conversion of LNA (18:3

37

n-3) to EPA (20:5 n-3) and DHA (22:6 n-3) (Kinsella 1990). Humans do not convert

LNA to 20-carbon fatty acids although taken in high doses (Dyerberg 1986).

For many years only n-6 PUFA LNA and AA were considered to be necessary

for normal growth and health, especially of the skin. An understanding of the

relationship between essentiality and bioconversion to physiologically active compounds

of such PUFAs has been clarified recently after the discovery and elucidation of the

synthesis and metabolic effects of eicosanoids (Weber et al. 1986).

Eicosanoids are metabolites of oxygenated derivatives of arachidonic acid and

other 20 carbon PUFAs (Weber et al. 1986). They are short-lived hormone-like

compounds produced by cells to participate in many physiological and pathological

processes such as blood clotting, stomach secretion and uterine contractions, and are

potent regulators of cell function (Weber et al. 1986, Turley and Strain 1993). The

eicosanoid family includes prostaglandins, thromboxanes, prostacyclins, leukotrienes,

lipoxins and other hydroxy fatty acids. Many functions of the circulatory, immune,

reproductive, secretory and digestive systems are modulated by prostaglandins,

thromboxanes and prostacyclins, while leukotrienes are involved in pulmonary functions,

inflammation and immune response (Pigott and Tucker 1990).

The substrate for eicosanoid production is commonly AA (C20:4 n-6) produced

from elongation and desaturation of LEA which is predominantly available in plant oils.

In the presence of n-3 PUFA, a series of eicosanoids is produced which displaces and/or

modifies the effects of those synthesised from n-6 fatty acids. Thus it results in a

decreased aggregation of blood platelets, with a reduction in thrombosis and ischaemic

heart disease and modification of immune functions, inflammatory and allergic reactions

(Kinsella 1987).

The dietary effects of n-3 PUFA are undoubtedly multifactorial. Beside

38

influencing the synthesis of eicosanoids, they can replace saturated fatty acids, decrease

synthesis of saturated fatty acids and decrease levels of circulating lipoproteins (Lands

1986). The concentration of serum TAG was lowered when n-3 PUFA was

incorporated into the diet and, at high doses, it has been shown to lower the content of

very low density lipoprotein (VLDL) cholesterol, which consequently lowered low

density lipoprotein (LDL) production, although there have been inconsistent effects on

LDL and high density lipoprotein (HDL) (Grundy 1986, Kinsella 1990, Sinclair 1993,

Turley and Strain 1993).

The reduction of plasma TAG and VLDL levels may be another beneficial

attribute of the n-3 PUFA, since the increase in serum TAG and cholesterol are

considered to be risk factors for coronary heart disease (Sinclair 1993, Turley and Strain

1993). Reduction of blood viscosity, systolic blood pressure and the extent of infiltration

of vascular cell walls by monocytes and neutrophils were also reported as biological

effects of dietary n-3 PUFAs (Kinsella 1990, Turley and Strain 1993).

2.3.3 Fish oil deterioration

Lipids, one of the components of flesh muscle, are unstable and degrade rapidly.

The most important changes in fish lipids are lipolysis, mainly involving enzymatic

hydrolysis, and autoxidation, mainly involving oxygen and unsaturated lipid (Hardy

1980, Huss 1988).

2.3.3.1 Lipolysis

During lipolysis, lipid undergoes hydrolysis producing free fatty acids and

glycerol as the major products. TAG and PL are the major classes of fish lipids, which

are hydrolysed by lipases and a group of phospholipases, respectively (Lovern and

Olley 1962).

39

The hydrolysis of TAG proceeds in steps via DAG, which are formed rapidly,

followed by the formation of MAG which takes place more slowly and is completed by

the very slow formation of glycerol (Brokerhoff and Jensen 1974). There are two kinds

of lipases reported to be present in fish: the lipases capable of hydrolysing long-chain

TAG such as those present in the red muscle of rainbow trout (Bilinski and Lau 1969),

and short-chain TAG lipases which is found in mackerel and lingcod (Shewfelt 1981).

Phospholipids, consisting of several fractions, are hydrolysed by a group of

phospholipases. Phospholipase A is commonly present in fish and its activity is

associated with the microsomal fraction (Shewfelt, McDonald and Hultin 1981, Han and

Liston 1987). Lysophospholipases have also been reported to be present in some fish

such as cod, rainbow trout and lizard fish and they are much more active than

phospholipases A (Shewfelt 1981, Sikorski and Kolakowska 1990). Phospholipases C

and D, which remove the nitrogenous base or the phosphate-base unit, are not

commonly found in fish muscle (Lovern and Olley 1962, Shewfelt 1981).

The activity of lipolytic enzymes in fish seems to depend on the species. Three

species of elasmobranch fish did not show any post-mortem hydrolysis of the flesh

phospholipid. On the other hand, in the gadoids and related species, the major FFA

content resulting from hydrolysis was caused by phospholipases, while in other species,

the phospholipases play an important role similar to lipases in producing FFA (Olley,

Pirie and Watson 1962, Addison, Ackman and Hingley 1969).

In a given species, the rate of lipolysis also depends on the type of muscle and

class of lipids. Bosund and Ganrot (1969) reported that the production of FFA in dark

muscle of Baltic herring during frozen storage was much faster than that in white muscle

and, in both types of muscles, PC was hydrolysed faster than PE. However, cod stored

in crushed ice showed a similar rate of hydrolysis both in PC and PE (Lovern, Olley and

40

Watson 1959). Ingenmansson, Kaufmann and Ekstrand (1993) reported that hydrolysis

of farmed rainbow trout (Oncorhynchus mykiss) during frozen storage resulted in

decreasing PC content of the muscle, losses ranging from 25-50%. In the case of PE,

the content remained constant in dark muscle, but decreased in white muscle.

The activity of hydrolytic enzymes also closely correlated with the time and

temperature of storage. Cod stored in crushed ice showed relatively low changes in the

initial phase, but the changes became markedly greater after about 2 weeks; 70% of the

phospholipid had been hydrolysed after about 5 weeks (Lovern, Olley and Watson

1959). Other studies of cod stored at various temperatures indicated that, at 0°C, there

was a marked initial lag phase before rapid hydrolysis of phospholipids began, while the

rate at 20°C was about 3 times that at 0°C (Olley and Lovern 1960).

During frozen storage, the hydrolysis of flesh lipids still continued, even at

temperatures below the freezing point. The rate reached a maximum level in cod stored

at - 14°C where the rate of hydrolysis was 10 times that at -22HC and 3 times that at 0°C

(Olley and Lovern 1960). In a separate study, the rate of hydrolysis of phospholipids in

cod flesh reached a maximum at about -4HC (Lovern and Olley 1960). de Koning and

Mol (1990) found that the rate of free fatty acid (FFA) formation in cape hake mince

stored at -5°C was much faster than that at -18°C and proceeded very slowly at -40°C.

Storage of cod fillets at -12°C for 9 months increased the FFA content from 5 to

326 mg/lOOg tissue with lipid content in tissue about 0.75% due to the hydrolysis of PE

and PC (Bligh and Scott 1966). The FFA in frozen cape hake mince stored at -18°C

gradually increased from 2.3% to 24.4% after 307 days of storage and this originated

both from phospholipids and neutral lipids (de Koning, Milkovitch and Mol 1987).

Srikar, Seshadari and Fazal (1989) observed that, during storage of marine catfish

(.Tachsurus dussumieri) at -20°C for 300 days, the FFA content increased from 1.9% to

41

5.4% resulting from hydrolysis of phospholipids.

Beside time and temperature of storage, processing also affected the hydrolysis of

fish lipid. Application of heat seemed to inhibit hydrolysis of lipid in fish such as cod

steamed for 90 min (Olley and Lovern 1960). The lipolysis in fish muscle was also

affected by dehydration, where lipases were active in a lower relative humidity

environment in freeze-dried halibut and tuna (Koizumi et al. 1978). The presence of

NaCl inhibited lipid hydrolysis in salted sardine fillet during storage (Takiguchi 1989) and

increasing the salt content from 2 to 8% in frozen sardine resulted in a gradually

decreasing rate of lipid hydrolysis (Nambudirv 1980).

2.3.3.2 Autoxidation

Fish oil containing fatty acids that are highly unsaturated are extremely

susceptible to autoxidative deterioration. The most important cause of deterioration in

the quality of fish oils from a flavour and odour standpoint is oxidation by atmospheric

oxygen.

As in the case for the oxidation of other fats, the first step of the oxidation

process is the formation of hydroperoxides involving a free radical chain reaction (Figure

2.3). Subsequently, the hydroperoxides react with oxygen to form secondary products

which either decompose into volatile breakdown products or condense into dimers

and polymers. In the presence of biological molecules in fish, such as pigments,

enzymes, and protein, the hydroperoxides and their breakdown products interact with

them resulting in discoloration and flavour and odour deterioration (Flsieh and Kinsella

1989).

There are many factors that affect oil oxidation in fish, including fatty acid

composition of lipids, disposition of the lipid within the tissue, the presence or absence

42

of activators and inhibitors (trace metals, haem compounds, oxidative enzymes,

antioxidants), and external factors such as storage time, storage temperature, light,

oxygen pressure, and aw (Flick, Hong and Knobl 1992, Hsieh and Kinsella 1989,

Lundberg 1967).

1. initiation

initiator

RH---------------- > R

2. propagationfast

a R* + 02----------- > ROOslow

b ROO* + RH--------- > ROOH + R*

3. termination• •

a. R + R ----------- > RR

b R + ROO*--------- > ROOR

• •

c ROO +ROO ------->R00R + 02

Figure 2.3 Scheme of lipid oxidation (Hsieh and Kinsella 1989)

Based on the fatty acid composition, the bond strength of saturated fats is high

and needs much energy for reaction. In unsaturated fats, hydrogen abstraction becomes

easier due to the presence of weaker carbon-hydrogen bonds especially in a doubly

allylic methylene system, therefore PUFAs such as AA, EPA and DHA containing 3, 4,

and 5 doubly allylic methylene systems, respectively, are much less stable than linoleic

43

and linolenic acids (Hardy 1980, Hsieh and Kinsella 1989).

Trace metals are very effective prooxidants and are active in the breakdown of

hydroperoxides (Tatum and Chow 1992). The relative activity of some metals and haem

compounds on lipid oxidation of fish flesh homogenate determined at 40°C was as

follows: Fe2 > haem compound > Cu2 > Fe3 (Hsieh and Kinsella 1989, Flick et al.

1992 ).

Haem iron, an important catalyst in oil oxidation, usually occurs in dark muscle,

therefore fish oils extracted from dark muscle may be more susceptible to oxidation than

that from white muscle (Ackman and Gunnlaugsdottir 1991, Hultin 1991).

Some enzymes that are commonly present in fish organs accelerate the oxidation

of PUFA. Peroxidase, lipoxygenase and the microsomal enzyme systems are reported as

the most important enzymes involved in lipid oxidation in fish (Han and Liston 1987,

Hsieh, German and Kinsella 1988, German, Zhang and Berger 1991). Peroxidase

isolated from fish leukocytes is capable of initiating lipid oxidation in the presence of

hydrogen peroxide and halides, while lipoxygenases that are able to initiate the

peroxidation of PUFA are commonly present in the gill and skin tissue of fish. The

microsomal enzyme system is another important enzyme that initiates lipid oxidation in

fish and other muscle foods.

It has been known that lipid oxidation is affected by temperature. For every 15

degree increase in the temperature above 60°C, the rate of oxidation of PUFA is

approximately doubled (Tatum and Chow 1992). The increasing temperature

accelerates not only the chain propagation reaction, but also peroxide decomposition

(Lundberg 1967). Time and temperature of storage seem to influence the oxidation of

fish lipid. Seer fish stored at -20°C showed an 8.4 fold increase in peroxide value after 6

months of storage (Fazal and Srikar 1987). The increase of PV and TBA number were

44

also found in fresh-water whitefish after 12 weeks of storage at -10°C (Awad, Powrie

and Fennema 1969). Ke et al. (1977) showed that the rate of peroxide formation in the

skin and the dark muscle of mackerel was significantly lower during storage at -40°C

than at -15°C.

Other catalytic factors, such as light, influence the rate of oil oxidation and the

accelerating effect of light is dependent on the wavelength (Lundberg 1967). Ultraviolet

light can provide the energy necessary for both initiation and propagation of oxidation

(Tatum and Chow 1992).

Oxygen partial pressure is an important factor of oil oxidation. Hsieh and

Kinsella (1989) reported that, at very low partial pressures, the rate of PUFA oxidation

is approximately proportional to the oxygen partial pressure, whereas at higher partial

pressure (>100 mm Hg), there is no dependence on the oxygen partial pressure.

The most important reason for deterioration in fish oil quality is autoxidation.

Undesirable flavours and odours develop at low PV values at an early stage of oxidation.

Hydroperoxides, the primary products of autoxidation, are tasteless and odourless.

However, they are very unstable, and oxidise further producing various short-chain

organic compounds, such as saturated and unsaturated aldehydes, ketones, acids and

other products, which contribute to the undesirable flavours and odours (Lundberg

1967, Kinsella 1987). The development of flavours and odours in oxidised fish oil

occurs at an early stage of oxidation and is described as "fishy", whereas at high levels of

oxidation a rancidity occurs, which is similar to that in other unsaturated oils (Lundberg

1967, Stansby 1990b).

Compared with other less unsaturated oils, the rapid increase in PV occurs earlier

in fish oils and is more rapid. When the PV is still low at an early stage of oxidation, the

odour and flavour may be described as similar to that of a good quality cod-liver oil, and

45

is quite different from that which develops later, which is generally described as fishy

(Stansby 1990b). This process is sometimes called "reversion", during which the

original flavour characteristic of the species gradually reappears, is followed by more

typical oxidation type flavours (O) and is later masked by the development of stronger-

flavoured substances referred to as "rancid". The degrees of fish oil rancidity are

distinguished, starting from very slight (the barely distinguishable rancidity with no after

taste) to the extremely rancid (exceedingly disagreeable flavour and after taste) as

presented in Table 2.7.

Table 2.7 The degrees of fish oil rancidity

Degree Flavourdescription

After taste description

very slight barely distinguishable usually absent

slight flavour weak but quite definite and somewhat objectionable to many

slight

moderate unpleasant to most lasting, undesirable to most

pronounced undesirable to almost everyone and quite objectionable

persistent, annoying

extreme exceedingly disagreeable very disagreeable

After Stansby (1990b)

The molecular compounds derived from the decomposition of oxidised fish oils

that contribute to odour and flavour reversion have recently been identified by many

researchers. A complex mixture of carbonyl compounds contributes to the fishy odours

and flavours with 2,4,7-decatrienal playing an important role in autoxidised oils

46

containing linolenic or n-3 fatty acids (Meijboom and Stroink 1972). Karahadian and

Lindsay (1989a) reported that the green and green-type flavours that initially develop in

the early oxidising fish oils are caused by trcms,cis-2,6 nonadienal, but trans-2 hexenal,

\,cis-5 octadiene-3-one and a low concentration of trans,trans,c/5-2,4,7-decatrienal add

heavier green, plant-like notes. Burnt/fishy flavours are caused by trans,cis,cis- and

trans,trans, c/s-2,4,7-decatrienal with c/s-4-heptenal as a modifier; while rancid and

painty flavours in fish oils were due to trans,cis-hexadienal, trans,trans-2,4 heptadienal

and trans, trans- and trans,cis-2,4 decadienal.

2.3.3.3 Treatments to prevent deterioration

It has been suggested that measurement of FFA content may be useful as an

index of degradation of nutritional components, since it has indirect effects on textural

changes by promotion of protein denaturation and on flavour deterioration by enhancing

lipid oxidation (Shewfelt 1981).

Since both autoxidation and lipolysis are the most important factors causing fish

deterioration and both are able to develop throughout catching, handling, storage and

processing of fish, it is necessary to prevent these processes in order to maintain the

quality of fish as a food raw material. Since the quality of fish products, such as fish oil,

is highly dependent on the quality of the raw material, maintenance of the freshness of

the raw material is the first priority before further processing is carried out. Since

deterioration is inevitable, it is requisite to get the fish to the market as soon as possible

or to submit it promptly to further processing.

The activity of endogenous enzymes is the main reason for the initial loss of

freshness, desiccation of the flesh, as well as oxidation of lipids and pigments. When

putrefactive bacteria are involved, undesirable changes in quality become more marked.

47

The rates of these processes are rapid and the effect of temperature is very important,

therefore deterioration should be retarded by cooling or chilling, which should be done as

soon as the fish is taken from the sea (Sikorski 1990).

Handling of the catch on-board is commonly by rapid chilling followed by

keeping them chilled during on-board storage until landed. In modem industrial fishing,

most vessels are equipped with mechanically refrigerated systems (Windsor and Barlow

1981, Kelman 1982, Bimbo 1990). In traditional fishing vessels commonly encountered

in less developed countries, preservation with ice is the most common way of chilling,

where the ice can be taken to sea and mixed with the fish manually. The quantity of ice

to be used is dependent on the insulation and refrigeration of the fishroom, the outside

air temperature and the duration of the trip, but 25-100% of the weight of catch is

generally used (Sikorski 1990), while 200% of the fish weight is ideal (Windsor and

Barlow 1981). In the case of long trip fishing, lowering the temperature of fish to

near the initial point of freezing tissue, i.e. -1.0 to -1.5° C is recommended to extend the

shelf life of fresh fish. Superchilling at -3°C to -4°C increases the shelf life of fish to 4-5

weeks, however, it is difficult in practical application.

Chilled sea water (CSW) made by mixing of sea water and ice in the proportion

of 1:2 (v/v) has a temperature of about -1.5°C and this can be used in small boats

(Sikorski 1990). Although the use of CSW for preservation has proved effective in

decreasing lipolysis in some fish, the salt in sea water can act as a prooxidant when

present at low concentration (Nambudiry 1980, Shewfelt 1981, Flick et al. 1992).

Evisceration reduces the risk of autolysis by digestive enzymes and eliminates

bacterial attack from the gut contents. This also serves for bleeding, thus reducing the

level of haem compounds and peroxidases commonly present in fish blood (Kanner and

Kinsella 1983), thereby reducing autoxidation potential. Gutting or eviscerating on

48

board is generally carried out only for white fish and large fish, while for small pelagic

fish and fatty fish it is impractical, especially when large hauls are taken (Kelman 1982,

Sikorski 1990). This situation, coupled with the intrinsically more rapid spoilage rate of

fatty fish, such as mackerel, herring and sardine, explains why, even under the best

conditions of handling and storage, these species have a much shorter shelf life than

gutted white fish such as cod. The best way to prevent the rapid autolysis and

autoxidation of fatty fish is to chill them rapidly as soon as they are caught, and to use

sufficient ice to keep their temperature at a low level for the duration of the fishing trip.

Sometimes the location of processing is far from the landing area. Therefore it is

important to bring such fish as quickly as possible to the processing area and to protect

them with sufficient ice during transportation . Upon arrival at the processing area, the

fish must be gutted as soon as possible and contamination of the fish with the gut should

be avoided. When gutting has finished, the gutted fish are washed immediately with

brine in order to remove the blood left in the tissue due to contamination during gutting

and bleeding.

When fish are not to be processed immediately on arrival at the plant,

preservation should be continued in order to retain the freshness of such fish. Freezing is

the oldest preservation method and is still commonly used worldwide. This is an effective

method of preventing microbiological spoilage and also can retard chemical and

biochemical deteriorative changes. Although freezing does not completely inactivate

most of the enzymes, the activity of these enzymes may be decreased significantly at

low temperature (Sikorski and Kolakowska 1990). In respect of lipolysis and

autoxidation, only low freezing temperatures (-30°C or lower) seem to slow such

processes significantly (Geromel and Montgomery 1980, Hardy 1980, Khayat and

Schwall 1983, de Koning and Mol 1990, Perez-Villarreal and Howgate 1991, Polvi et

49

al. 1991). At temperatures only a few degrees below freezing (-4°C to-14°C) the rate

of hydrolysis increases (Olley and Lovern 1960, Lovern and Olley 1962). Most modern

freezers such as air blast, plate, and cryogenic freezers are capable of cooling

unpackaged fish from 5°C to -30°C in 4 h or less (Graham 1982). Glazing with a thin

layer of ice helps to minimise the dehydration of unwrapped frozen fish, especially for

long term storage (Hardy 1980, Sikorski and Kolakowska 1990, Flick et al. 1992).

Vacuum packaging or controlled atmosphere packaging is effective, especially

when it is combined with frozen storage. Vacuum packaging at -20°C inhibited

hydrolytic rancidity of menhaden mince (Hwang and Regenstein 1988), while the

storage life of Atlantic mackerel was extended by vacuum packaging combined with

frozen storage at -26°C (Flick et al. 1992). The use of laminated plastic packaging

materials coupled with vacuum packaging has proved effective in preventing rancidity in

mackerel and herring fillets (Bligh and Merritt 1988).

Since heat and oxygen enhance oil oxidation, minimising heating and contact with

oxygen during extraction of oil is necessary. The use of equipment made from material

easily corroded, especially cuprous metals, is strictly forbidden in any oil processing plant

and the use of stainless steel or tin-plated metals is recommended. Beside extraction,

further processing of fish oil must be carried out in an appropriate manner. According to

Tatum and Chow (1992), air exclusion and the use of a hermetic centrifuge in the

refining process are recommended during degumming and neutralisation, while in

physical refining and deodorisation, both oil and the steam should be fully deaerated and

the vessel used must be completely leak free. After processes that apply heat treatment,

the oil should be cooled immediately and stored at a low temperature in inert airtight

containers and protected from light. The equipment used for refining and storage of fish

oils must be made from non-corrosive metals to avoid contact of fish oil with metal

50

catalysts.

The use of antioxidants is commonly necessary to minimise fish oil oxidation,

especially during storage. An antioxidant is a substance that inhibits the initiation steps

or the early propagation steps (Olcott 1967). The antioxidant free radicals combine

with those of the autoxidation process to give stable non-radical compounds, as a

consequence, radicals are removed from the system and the chain reactions associated

with autoxidation are broken. Figure 2.4 shows the basic mechanism of antioxidant

(AH) inhibition.

H H H H H HII I III

R-C-C=C-R1 + AH ------------- > R-C-C=C-R1 + A*

H

H H H H H H

R-C-C=C-R1 + AH................... > R-C-C=C-R1 + A'

O O

O O

H

H HH

A* + R-C=C-C-R1 (does not react)

A' + A' ■> AA (terminates)

Figure 2.4 Mechanism of antioxidant inhibition (Olcott 1967)

51

The natural antioxidant commonly present in most fish oils is tocopherol (Kinsella

1987, Stansby 1990b), ranging from about 0.2 to 270 mg/100 g edible portion (Sikorski

et al. 1990b). However, losses usually occur due to oxidation during preparation and

processing (Young 1986, Sikorski et al. 1990b, Tatum and Chow 1992), thus the

addition of synthetic antioxidants is necessary. Antioxidants should be carefully used and

not all are applicable to certain seafoods (Flick et al. 1992).

Studies on the use of synthetic antioxidants to control lipid oxidation in fish oil

have been reported. Thorisson, Gunstone and Hardy (1992) found that ethoxyquin,

which is extensively used in fish meal production, was effective for fish oil at the level of

0.1% and 0.5%. TBHQ at 0.01% was the most powerful antioxidant compared to

Anoxomer™ (a synthetic nonabsorbable polymer with a phenolic nature), ethoxyquin,

butylated hydroxytoluene (BHT) and BHA for (capelin) fish oil (Kaitaranta 1992). The

TBHQ (0.02%) was also found more effective in inhibiting the oxidation of mackerel

skin lipid when compared to other antioxidants such as a-tocopherol, tempe (an

Indonesian traditionally fermented soybean) oil, BHA and BHT (Ke, Nash and Ackman

1977). On the contrary, Hwang and Regenstein (1988) reported that ascorbic acid and

erythorbic acid had better antioxidant activities than tocopherol, rosemary extracts and

TBHQ. Indeed, TBHQ had only limited activity. Several flavonoids, either commercial

(catechin, morin and quercetin) or natural (5,3',4'-trihydroxy-7-methoxy flavonone

designated as Pt-2), could be used as natural antioxidants and might substitute for

synthetic antioxidants such as BHA and BHT (Nieto et al. 1993).

2.3.4 Methods for lipid extraction

Lipids are present in living tissues both in the free form; commonly the neutral

lipids and bound in membranes in forms such as the lipoproteins; commonly the more

52

polar lipids, such as the PL (Kates 1986). For that reason, it needs solvents or solvent

mixtures that are sufficiently polar to remove all lipids from their association with cell

membranes or lipoproteins, but should not react chemically with the lipid and should not

be so polar that TAGs and other non-polar simple lipids do not dissolve and are left in

the tissue (Christie 1982, Nelson 1991).

Single pure solvents are not suitable as general purpose lipid extractants. Non­

polar solvents will not extract the polar lipids from tissues under most circumstances,

while polar solvents alone will leave non-polar lipids in the residue and require several

solvent changes, resulting in a slow extraction process (Nelson 1991).

2.3.4.1 Common extraction methods using single or multiple solvents

Many extraction methods using either single or multiple solvents have been

applied to various tissue matrices. Some single solvents commonly used in lipid

extraction are acetone, diethyl ether, isopropanol and hexane.

Stansby (1951) reported that hot acetone was superior to diethyl ether in

extracting lipid when refluxed with fish meal, and when acetone was mixed with 12N

hydrochloric acid (1% v/v), extraction was much more rapid. However, by comparison

with diethyl ether, acetone extracted considerably more non-lipid material from fish meal.

In addition, the extract was not suitable for determination of the deteriorative changes

in the original lipids, such as FFA or hydroperoxides (Lovern 1965).

Soxhlet extraction employing petroleum ether or diethyl ether was reported to be

unsatisfactory for lipid composition determination in meats (Sahasrabudhe and

Smallbone 1983), while 4N HC1 digestion, followed by diethyl ether extraction, was the

most effective method for extracting lipids from food products (Sheppard, Hubbard and

Prosser 1974).

53

The use of dry isopropanol to extract fat from fish material was also reported to

be more efficient than dry acetone, however, the isopropanol extract also contained a

considerable amount of water soluble material and required more thorough purification

than the fat extracted by acetone (Dambergs 1959).

de Koning et al. (1985) stated that hot hexane, applying either a reflux or the

Soxhlet method, did not completely extract lipid bound to protein, and digestion of the

sample using 3M HC1 before hot hexane Soxhlet extraction resulted in considerable

hydrolysis of the PL. Thus a correction factor was needed to account for the losses of

some lipid in the hydrolysis step.

Different combinations of solvents have also been applied to extract lipid from

different tissue matrices. Entenman (1957) suggested that a mixture of either alcohol-

diethyl ether (3:1 v/v) or alcohol-acetone (1:1 v/v) applying hot extraction was excellent

to extract lipid from body fluids such as serum and plasma, while treatment with hot

alcohol combined with extraction using either diethyl ether or chloroform in a Soxhlet

apparatus was useful for large tissue samples and from the whole bodies of small animals.

Soft tissue such as liver could be extracted using chloroform in a Soxhlet apparatus after

grinding it with anhydrous sodium sulphate.

Unlike the methods above which commonly apply heat during extraction, Folch

et al. (1951, 1957) and Bligh and Dyer (1959) introduced wet extraction methods using

a chloroform-methanol system. These techniques offer the advantages of short

processing times and avoid the necessity of applying heat to effect the extraction.

Sahasrabudhe and Smallbone (1983) stated that a mixture of methylene chloride-

methanol (2:1 v/v) or chloroform-methanol (2:1 v/v) employing Soxhlet extraction was

effective in extracting lipid from beef and reported it to be comparable to the wet

extraction method employing similar solvents as introduced by Folch et al. (1957) and

54

Bligh and Dyer (1959). The chloroform-methanol system will be discussed further in

Section 2.3.4.2.

Since fish lipids contain a considerable amount of fatty acids that are highly

unsaturated, an ambient temperature wet extraction method is preferred. Erickson

(1993) compared different solvent systems in extracting lipid from channel catfish

muscle. The author reported that mixtures of chloroform-methanol (2:1 v/v), hexane-

isopropanol (3:1 v/v), chloroform-isopropanol (7:11 v/v), dichloromethane-methanol

(2:1 v/v) and chloroform-methanol-water (2:2:1 v/v/v) were similar in recovery of both

TAG and PL, whereas a mixture of sodium dodecyl sulphate (SDS)-ethanol-hexane

(either 4:10:3 or4:10:9 v/v/v) failed to extract PL from muscle tissue.

2.3.4.2 Chloroform-methanol solvent system

The use of chloroform and methanol-based systems has been recognised as

suitable for lipid extraction in animal tissue (Folch et al. 1951, Bligh and Dyer (1959).

Folch et al. (1951) introduced a simple extraction method for brain tissue by

homogenising the tissue with a mixture of chloroform-methanol (2:1 v/v) for about 3

min. Insoluble matter was removed by filtration and the filtrate was washed with water

(5-fold volume) to remove non-lipid contaminants. The author noted that 1% of lipid

was also removed during washing, and after modification of the method (Folch et al.

1957), in which the extract was washed with 0.2 volume of either water or an

appropriate salt solution, the lipid losses were reduced to only 0.3-0.6%. This method

was reported to be simple and applicable to any scale, gave low losses of lipid in the

washing process and yielded a washed extract that could be taken to dryness without

foaming and without splitting of the proteolipids (Folch et al. 1957)

The Folch method actually is a tertiary solvent system consisting of chloroform-

55

methanol-water. This was further explored by Bligh and Dyer (1959) who developed the

chloroform-methanol-based extraction system by emphasising the important role of

water in the extraction of lipid from tissue and the consequences of the solvent-to-water

ratios on subsequent purification of the total lipid extract. In the Bligh and Dyer (1959)

method, which was applied to wet fish, almost complete lipid extraction could be

achieved by homogenising the sample in a one-phase solvent system, using chloroform-

methanol-water (1:2:0.8 v/v/v) in which the water in the original sample was taken into

account. Additional chloroform and water were then added to form a biphasic system,

so that the composition of the chloroform-methanol-water ratio became 2:2:1.8 (v/v/v).

When these solvent ratios are observed, the lipids remain in the chloroform phase, while

the non-lipid substances concentrate in the methanol-water phase. The losses of lipid in

the Bligh and Dyer (1959) method were also 1%, similar to those of the original Folch

method. To achieve best results, the proportion of the solvents both in the one phase

system (before dilution) and two phases (after dilution) should be followed precisely

since, when the proportion is shifted toward either or both more water or methanol, the

Bligh and Dyer (1959) method will not remove the majority of the impurities from the

extract (Nelson 1975). For that reason, the water content of the original sample should

be measured before extraction.

Regarding solvent-sample ratio, Folch et a/. (1957) recommended the ratio of

solvent to sample of 20:1 (v/w) , while in the Bligh and Dyer (1959) method, the final

solvent to sample ratio is only 5.5:1 (v/w). The important improvement of the Bligh and

Dyer (1959) method lies in the great reduction of solvent-sample ratio and in

simplification and speeding up of the entire operation.

Although the use of chloroform and methanol presents major environmental

hazards, the method is simple and rapid since it combines extraction and purification into

56

a single step and is effective in extracting lipid without lipid degradation; therefore such a

method is suitable for extraction of lipid from fish containing highly unsaturated fatty

acids (HUFA). However, the aqueous washing procedure applied in this method will

remove most of the gangliosides and ceramide polyhexosides in the original extract.

Acidic phospholipids partly retained in the aqueous phase can also be lost (Palmer 1971,

Kolarovic and Fournier 1986). Although Bligh and Dyer (1959) noted that the amount

of lipid loss due to the washing procedure was only 1% of the total lipid, the losses of

lipid will be higher when lipid is extracted from fish that has been cold-stored for a long

time, due to the dissolution of enzymatic hydrolysis products (fatty acids and

phosphatidic acids) to the methanol-water phase, which is discarded (Olley and Lovem

1960, Hardy, McGill and Gunstone 1979, Joseph and Seaborn 1982).

To prevent inaccuracies of filtration and re-extraction of the residue, Hanson and

Olley (1963) modified the Bligh and Dyer (1959) method by centrifugation of the whole

mixture and an aliquot was taken from the lower layer. They suggested that the water

content of the sample should be adjusted so that the correct solvent proportion is

achieved. This method proved satisfactory with 1-20 g of fish. Erickson (1993) also

suggested that the losses of PL can be prevented when using a solvent system of

chloroform-methanol-water by separating the extract from the tissue residue prior to

formation of a two-phase system.

2.3.4.3 Hexane-isopropanol solvent system

A single phase hexane and isopropanol solvent system is one of the extraction

methods that is considered to be able to overcome the problems encountered in the Bligh

and Dyer (1959) method (Kolarovic and Fournier 1986, Joseph and Seaborn 1990).

This extraction system was introduced by Hara and Radin (1978) when it was applied to

57

the extraction of lipid from rat or mouse brain. In this method, one part of brain was

homogenised in 18 parts of a mixture of hexane-isopropanol (3:2 v/v) for 30 sec,

followed by filtration of the suspension to separate the aliquot containing lipid from the

residue. The residue was then re-suspended in 2-3 portions of similar solvent, and

allowed to soak for 2 min. before filtration. Hara and Radin (1978) noted that the

hexane-isopropanol extract can be processed without washing in most cases, however,

when necessary, the extract can be washed with aqueous sodium sulphate to remove

non-lipids.

The solvent system used in the above method consists of a low-polarity solvent

(hexane) and a high polarity, water miscible solvent (isopropanol), which can penetrate

cell membranes and dissolve a wide range of lipids that differ markedly in their solubility

(Radin 1981). In addition, it extracts almost no protein, less pigment and non-lipid

material from tissues. However, it extracts only a part of the gangliosides sufficiently.

2.3.5 Analytical methods to determine fish lipid stability and flavour

2.3.5.1 Lipid stability

Many methods are used to assess lipid stability, including static and dynamic

methods. The static methods involves PV, thiobarbituric acid reactive substances

(TBARS), polyene index (PI) and fluorescence detection, while weight gain techniques

and induction period determined by a Rancimat are categorised as dynamic methods.

2.3.5.1.1 Peroxide Value (PV)

The first product formed during lipid oxidation is a hydroperoxide, therefore

measurement of the concentration of hydroperoxide can be used to evaluate the content

of primary oxidation products.

58

The common method for measuring the concentration of hydroperoxide is PV

determination applying an iodometric technique, which is reported as milliequivalents

(meq) of iodine per kilogram of fat (Rossell 1989). Since peroxides are vulnerable to

further reaction, the estimation of hydroperoxides by P V indicates the oxidation status of

an oil only in the early stages of oxidation. As oxidation continues, the hydroperoxides

decompose into low molecular weight volatile compounds such as aldehydes, ketones,

alcohols, acids and hydrocarbons resulting in under estimation of the degree of oxidation

(Gray 1978, Gray and Monahan 1992).

The use of the PV method to assess the quality of highly unsaturated oils, such as

unhydrogenated fish oils, is limited. This might be due to the quick conversion of the

hydroperoxide initially formed into secondary oxidation products, resulting in a lower

amount of peroxide oxygen relative to that in vegetable oils or animal fats, even when

these oils have undergone extensive oxidation (Rossell 1989).

The size of sample used in this method is relatively high, i.e. 5g/replicate analysis

and this is not a problem when it is applied to the analysis of oxidised rendered oils. It

will be a problem when oxidised lipid in fish or other animal tissue is to be determined,

since large quantities of tissue must be extracted to obtain sufficient lipid for analysis

(Joseph and Seaborn 1982). Several spectrophotometric PV measurement methods

which appear promising have also been suggested (Hicks and Gebicki 1979, Asakawa

and Matsushita 1980).

2.3.5.1.2 Thiobarbituric acid reactive substances (TBARS)

The TBARS test is one of the analytical methods widely used for the

determination of lipid oxidation. The extent of lipid oxidation is reported as TBARS

number or TBARS value which is expressed as mg of malonaldehyde per kg tissue or per

59

kg oil (Gray and Monahan 1992, Melton 1983).

The method is based on the reaction of TBARS with thiobarbituric acid to give a

red chromogen that may be determined spectrophotometrically. The coloured complex

produced has an absorption maximum at 530-532 nm (Gray and Monahan 1992). The

chromogen is formed through the condensation of two molecules of TBA with one

molecule of malonaldehyde (Sinnhuber, Yu and Chang 1958), however, the evidence

that malonaldehyde could be found in all oxidising systems is still questionable (Gray

1978).

The reaction of TBA with other products of lipid oxidation, such as various

classes of aldehydes, also produces colour complexes having an absorption maximum

similar to that of the malonaldehyde-TBARS complex (Marcuse and Johansson 1973).

Besides, TBARS can also react with compounds other than those found in oxidising

systems, such as certain carbohydrates (Baumgartner et al. 1975, Dugan 1955) and the

decomposition products of many amino acids (Buttkus and Bose 1972).

Dahle, Hill and Holman (1962) reported that the amount of malonaldehyde

produced in an oxidising lipid appeared to depend on the degree of unsaturation of the

fatty acid component, whereby the molar yield of the TBARS colour increased with the

degree of unsaturation. Only peroxide radicals with an ethylenic bond located beta and

gamma to the peroxide group were capable of undergoing cyclisation with the formation

of malonaldehyde and thus the requisite chromogen could be produced only from fatty

acids containing three or more double bonds. Therefore tissue samples with different

fatty acid profiles result in different values of TBARS (Joseph and Seaborn 1982). Thus

the TBARS test can give meaningful results only when it is applied to a comparison of

different oxidation stages of samples from a single material.

Pryor, Stanley and Blair (1976) confirmed that malonaldehyde was formed from

60

oxidised trienes, but not from dienes, while other authors suggested that TBA reagent

reacts with malonaldehyde resulting from autoxidising fatty acids with less than three

double bonds (Tarladgis et al. 1960, de Koning and Silk 1963, Castell et al. 1966,

Ohkawa, Ohishi and Yogi 1978, Caldironi and Bazan 1982).

The TBARS test can be performed on the whole sample by directly heating the

food samples with TBA solution and extracting the red pigment, on a portion of the

steam distillate of samples, on aqueous acid extracts of samples and on the extracted

lipid from samples. The absorbance of the red pigment produced from these reactions

with TBA is measured spectrophotometrically (Hoyland and Taylor 1991, Rahardjo and

Sofos 1993).

The direct heating of the whole sample in TBA solution is reported to enhance

the degradation of the sample itself (Tarladgis et al. 1960), resulting in an overestimation

of the malonaldehyde content of the sample before analysis. Interfering substances such

as protein, amino acids and nucleic acids may also react with TBA, resulting in difficulty

of interpretation of the TBA number (Rahardjo and Sofos 1993).

The distillation method has been found to minimise direct interference by non­

lipid components on the TBA reaction, since there is no direct contact between sample

and the TBA, and the clear aqueous solution of TBARS obtained is easier to measure

(Tarladgis et al. 1960). According to William et al. (1983), this method is more

sensitive and more suitable for high fat samples. A modification of this method is

reported to be simple, sensitive and satisfactory when used for fish flesh (Ke, Cervantes

and Robles-Martines 1984). However, the main disadvantage of this method is that

distillation is an empirical procedure requiring the collection of a specific volume of

distillate (Hoyland and Taylor 1991) and it is also time consuming.

The estimation of the malonaldehyde content after aqueous acid extraction does

61

not involve heat treatment, hence this method has an advantage over the two methods

previously discussed. Witte, Krause and Bailey (1979) noted that the solvent extraction

method is easier to use than the distillation method, uses less equipment and heating is

not essential. A modification of this method can decrease the amount of interfering

substances and can thus more specifically measure only the malonaldehyde (Rahardjo and

Sofos 1993).

The measurement of TBARS on the lipids extracted from a sample is reported to

be faster and easier than distillation, and is recommended for analysis of a large number

of samples needing to be analysed rapidly (Pikul, Leszczynski and Kummerow 1989).

This method also gives a higher recovery when compared to the distillation method

(William et al. 1983, Pikul et al. 1989). The advantage of this method is that the

presence of interfering substances can be eliminated and Pikul et al. (1989)

recommended this procedure as particularly appropriate to study the susceptibility to

oxidation of different kinds of lipids or individual lipid components. In spite of these

advantages, the recovery of malonaldehyde might be low since malonaldehyde is mostly

present in the aqueous phase of the sample and only a minor portion of malonaldehyde is

present in the organic phase together with the extracted lipids (Schmedes and Holmer

1989). In addition, evaporation of the lipid fraction in the organic phase gives a greater

chance for the lipids to oxidise (Rahardjo and Sofos 1993).

Although the use of more sensitive equipment has been proposed, such as

spectrophotofluorometry (Sawicki, Stanley and Johnson 1963), GC (Hamberg, Niehaus

and Samuelsson 1968) and HPLC (Kakuda, Stanley and Van de Voort 1981), there are

still limitations of the TBARS test, i.e. unsuitable to analyse samples where the products

of lipid oxidation are irreversibly bound to macromolecules, and for following lipid

oxidation in some shelf life studies where TBARS values fluctuate as oxidation

62

progresses (Hoyland and Taylor 1991).

2.3.5.1.3 Polyene index (PI)

Lipids that contains PUFA, such as those found in marine products, are prone to

undergo oxidation resulting in changes in the fatty acid composition. May and McCay

(1968) used the losses of PUFA as an index to detect and measure the extent of lipid

peroxidation. They observed that PUFA, specially AA in liver microsomes of rats,

disappeared as a result of oxidation and suggested that such PUFA were converted to

some unidentified moiety on the phospholipid. Keller and Kinsella (1973) also

suggested that changes in fatty acid composition of lipids provide an indirect measure of

susceptibility to lipid oxidation. Shono and Toyomizu (1972) used the decrease of the

ratio of C22.6 to Ci6:o as an index of oxidative rancidity of lipids in fish products.

However, since in some fish oils C20:5 n-3 may be considerably more important than C22:6,

Ke, Ackman and Linke (1975) suggested the use of the ratio of total polyenoic acids

(C 18:4 + C20:5 + C22:6) to total saturated acids (C 14:0 + Ci6:o + Ci8:o) as PI related to the

concentration of 2,4,7-decatrienal causing fishy flavour in rancid mackerel oil. The

authors reported that the ratio of the formation of 2,4,7-decatrienals was roughly

proportional to the increase of PV and decrease of PI in the highly oxidised mackerel oil,

and the polyenoic fatty acids were reported to decrease markedly from 25 to 13 molar

percent, while monoenoic fatty acids changed only slightly, i.e. from 43 to 37 molar

percent, presumably due to oxidation.

Changes in the levels of fatty acids and PI in the lipids from different parts of

mackerel during storage were noted by Ke et a/. (1977), where the proportion of

saturated acids increased while the proportion of PUFA decreased. In the skin lipids of

oil sardine during storage at -18°C, the PI significantly decreased in the initial stage,

63

indicating a higher rate of development of rancidity, but after storage for 22 weeks, the

changes of PI were faster in muscle lipids (Khayat and Schwall 1983).

Suzuki et al. (1985) reported that, within one month storage at 22°C, the levels

of C22:5 and C22.6 fatty acids decreased rapidly and became stable thereafter, while during

storage at 2°C the levels of both C22:5 and C22:6 fatty acids decreased slowly. When

stored at -30°C, no significant changes of these acids were noted. Hardy et al. (1979)

found that a slight, but relatively consistent loss of the C20 and C22 acids, such as DHA,

was detectable in cold stored (-10°C) cod samples. On the contrary, Polvi et al. (1991)

found that n-3 fatty acids in Atlantic salmon muscle were relatively stable even under

adverse storage conditions and during frozen storage at -12°C. Very little loss of TAG

and no selective changes in TAG fatty acid composition were noted after 3 months

storage.

In the processing of dried salted marine catfish, a 30% loss of PUFA was found

during salting, but during drying and storage, no subsequent loss was noted (Smith et al.

1990). Lubis and Buckle (1990) stated that PI (C20:5 + Cjj JCxe-.o) values measured from

dried salted sardine stored at 5°C, 20°C and 30°C decreased during storage and this index

showed a strong correlation with the rancidity score (TBARS) of samples.

Determination of the loss of polyunsaturation of an oil by GC of the fatty acid

constituents appears to be the most direct means of evaluating the extent of lipid

oxidation. However, it is essential that the original fatty acid composition of the

unoxidised lipids be known when this method is applied (Joseph and Seaborne 1982).

In addition, the authors found that loss of Ci6:o was encountered during the oxidation of

some marine lipids, which might be due to intramolecular bonding of fatty acid

constituents or possibly development of polymers involving unesterified 16:0, therefore,

they suggested that the use of native saturated fatty acids (e.g. Ci6:o) as internal standards

64

is not advisable.

2.3.5.1.4 Fluorescent products

One of the major secondary products of lipid peroxidation is malonaldehyde.

This compound reacts with amino groups in amino acids to form fluorescent, conjugated

Schiff base products with the base structure R-N=C-C=C-N-R having emission maxima

at 460-470 nm when excited at 360-400 nm (Ohio and Tappel 1969). Compounds with

similar fluorescent spectral characteristics are also produced from the interaction of

peroxidising lipids and cellular constituents, such as phospholipids (Bidlack and Tappel

1973, Dillard and Tappel 1973), amino acids (Trombley and Tappel 1975, Shimasaki,

Privett and Hara 1977), proteins (Buttkus 1967) and deoxyribonucleic acid or DNA

(Reiss and Tappel 1973). Buttkus (1975) reported that the formation of fluorescent

compounds is due not only to the malonaldehyde-amine reaction, but might also be due

to self-condensation and polymerisation of malonaldehyde having a fluorescence

spectrum different from that of malonaldehyde-amine conjugates. Extracts of rancid

herring had fluorescence spectral characteristics of both malonaldehyde-amine and

malonaldehyde self-condensation products.

Fletcher, Dillard and Tappel (1973) noted that the major portion of fluorescent

molecular damage in biological tissue is found in the lipid-soluble phase of the

chloroform-methanol extracts. These compounds, obtained from peroxidation of tissue

containing unsaturated fatty acids with 2, 3 and 4 double bonds, showed fluorescence

with excitation maxima in the 345-360 nm region and emission maxima in the 425-450

nm region, while samples having appreciable amounts of 4, 5 and 6 double bonds were

characterised by fluorescence with excitation at 360-380 nm and emission at 440-470

nm.

65

Fluorescence measurement has been developed and successfully used to

determine lipid peroxidation damage of biological materials (Fletcher et al. 1973, Dillard

and Tappel 1971) and the increase in fluorescence correlates well with the increase in

oxygen absorption and TBARS production (Dillard and Tappel 1971). Lubis and Buckle

(1990) used fluorescent products as one of the indices to evaluate lipid degradation in

dried-salted sardine during storage. They found that, as with the case of PI, fluorescent

products also showed significant correlation with the rancidity score.

Although detection of malonaldehyde in the TBARS test is the assay most

frequently used for assessing lipid oxidation in muscle foods and other biological

systems, one problem encountered in using the method is that several factors can also

affect the intensity of the colour complex due to reaction of TBA with other products

that are not oxidation products, e.g. glyceraldehyde, hydroxy methylfurfural, sucrose and

woodsmoke components, to form a red complex that has the same absorption maximum

as the malonaldehyde-TBA complex (Gray 1978, Gray and Monahan 1992).

Since malonaldehyde is a very reactive material and the reaction of

malonaldehyde with protein and other cellular constituents results in fluorescent

products, the detection of lipid oxidation in biological tissue using the fluorescence

technique has been reported to be 10-100 times more sensitive than that using the

colorimetric TBA test (Dillard and Tappel 1971). Moreover, the development of

fluorescent products is closely parallel to oxygen absorption even in the later stage of

peroxidation.

Although the fluorescence method is very sensitive, the chemistry of the

formation of fluorescent products and their characterisation needs further exploration. In

addition, the application of this method in assessing lipid oxidation from different

products needs to be investigated.

66

2.3.5.1.5 Weight gain

This method is based on gravimetric monitoring of the weight increase due to the

oxygen absorption of the sample heated in an oven. Although this method has some

disadvantages, such as the possibility of an irreproducible result due to discontinuous

heating of the sample during weighing, tedious work, especially when a high frequency

of weighing is required, and error due to non-standard conditions (shape of container,

amount and size of sample and temperature), the method has a low instrumentation cost

and no limit to the number of samples to be processed (Garcia-Mesa, Luque de Castoo

and Valcarcel 1993). The use of a small amount of sample to keep a large

surface/volume ratio and small container with test tube diameter, coupled with high

weighing frequency with strict timing of the weighing process, can create optimal

working conditions (Garcia-Mesa et al. 1993).

This simple method was used to assess the effect of some metal catalysts on the

oxidation of lipids in mackerel skin and meat at 60°C (Ke and Ackman 1976) and an

appropriate linear relationship was obtained between weight gain and other parameters

of lipid oxidation such as PV, volatile carbonyls, polyene and monoene indices. A

comparative study on the activity of several antioxidants on the oxidation rate of skin

and meat lipids of mackerel was also carried out by determining the time at 60°C

required to reach the end of the induction period and to a 1% sample weight increase

(Ke et al. 1977), while Kaitaranta (1992) determined the efficiency of various

antioxidants to control lipid oxidation in fish oil based on a 0.6% weight increase of the

oil. This weight increase was taken as corresponding to a 27% content of polar

artefacts in frying oil.

The weight increase was also reported to have a direct linear correlation with the

increase in levels of polar lipids and a hyperbolic correlation with the TBARS value

67

during an accelerated oxidation test of fish oils (Kaitaranta and Ke 1981).

2.3.5.1.6 Rancimat method

Traditionally, the active oxygen method (AOM) has been used to determine

oxidative stability, in which the sample is exposed to a stream of dry air at a temperature

of 97.8°C, and periodically samples of the oil are withdrawn and the PV is determined

(Ragnarsson and Labuza 1977). When PV is plotted against time, a characteristic curve

is obtained comprising an induction phase, in which practically no hydroperoxides are

formed, followed by an oxidation phase, during which a large increase in PV is detected.

The time required to achieve a large increase in PV, recognised as the induction time,

has been used as a good indication for keeping properties and quality status of fats or

fatty products. Fresher fats generally have longer induction times. The end point of

induction time varies from oil to oil depending on its stability.

To simplify this method and to eliminate the disadvantage of its labour intensive

nature, equipment such as the Rancimat has been developed. The Rancimat is an

automated method analogous to the AOM developed by Hadorn and Zurcher, which is

based on the observation that, in an oxidising oil, volatile acids are formed at the end of

the induction period. These volatile components are trapped in distilled water, measured

conductimetrically and the conductivity plotted automatically (Laubli and Bruttel 1986).

de Man, Tie and de Man (1987) noted that the volatile acids produced by several

oils were composed mainly of formic acid with a significant amount of acetic acid. Other

acids with three or more carbon atoms such as propionic, butyric and caproic, were also

produced in small amounts. The relationship between PV, conductivity and the organic

acids content was also established.

Measurement of oil stability assessed by the Rancimat is commonly carried out

68

using a 2.5g sample, with an air flow rate of 20L/h, held at 110-112°C (Frank, Geil and

Freaso 1982). Flowever, selection of a suitable temperature allowing the induction time

of fats and oils which are either very stable or less stable toward oxidation is suggested in

order to obtain a reasonable range of induction period (5-20 h). A very short induction

time should be avoided to prevent a great scatter in the results (Laubli and Bruttel 1986).

According to Frank et al. (1982), the induction time seems not to be affected

significantly by changes both in sample size and air flow rate, but the shape of the graph

is affected. Increase in air flow tends to increase the steepness of the curve both before

and after the induction time.

When several oils and fats were analysed by the standard AOM and by Rancimat,

the induction times correlated well (Laubli and Bruttel 1986). Measurement of the

stability of edible oils in the presence of antioxidants and prooxidants was also

performed by Rancimat (Frank et al. 1982). Eyers (1991) evaluated with the Rancimat

the changes in oxidative stability of sunflower oil, that was stored either at ambient or

elevated temperatures, and the results correlated well with sensory evaluation. However,

his finding on the evaluation of antioxidant effectiveness in fresh oil and oil stored at

ambient temperature, did not correspond with either chemical or sensory evaluation.

Since the slightest traces of metals or of certain other substances from the sample

container can have a very marked effect on the induction time, perfectly clean reaction

vessels are absolutely essential to obtain reliable results (Frank et al. 1982, Rossell

1989). The advantage of using the Rancimat over AOM is less labour and less expense

due to the absence of continual PV determination. Moreover, no supervision is required

during measurement (Laubli and Bruttel 1986).

69

2.3.5.1.7 Headspace analysis

Headspace analysis is becoming a popular method to monitor the lipid oxidation

of oils and foods containing oil. This method is based on measurement and

characterisation of the volatile secondary products resulting from lipid oxidation that are

measured directly by gas chromatography (Prior and Loliger 1994). The advantage of

this method is that it can measure the concentrations of the volatile components that

constitute the rancid odour. The measurement of the volatiles can be performed by

several techniques such as static headspace and dynamic headspace, steam

distillation/extraction, and these are discussed in Section 2.3.6.

2.3.5.2 Flavour

Flavour and aroma are the most important factors reflecting the quality of fish

and fish products, and are directly related to consumer acceptance. Therefore, the most

reliable method reflecting the characteristic fish flavour and aroma is sensory assessment.

However, it is time consuming, and importantly the assessor is subject to fatigue

resulting in erroneous assessment (Josephson, Lindsay and Olafsdottir 1987). For that

reason, objective methods to measure fish flavour and aroma are necessary.

The presence of flavour compounds at low concentration in food matrices,

complexity of aroma volatiles and extreme volatility and instability are factors

contributing to the difficulties in analysing flavour quality. However, advances in

analytical methodology have been helpful in trying to overcome these problems.

Flavour analysis commonly consists of isolation and concentration of the aroma

compounds, separation of aroma components, odour characterisation of individual

flavour compounds and identification of the individual volatile compounds.

70

2.3.5.2.1 Isolation and concentration of aroma compounds

In the investigation of volatile food components, isolation of volatile constituents

contributing to characteristic aroma is the first step. Before isolation, special treatments

including crushing, blending, mincing and grinding are applied to extract the flavour from

the food matrices (Reineccius 1994).

Flavour isolations have been performed by many different techniques including

direct injection (static headspace), dynamic headspace sampling (DHS) or headspace

concentration, distillation/extraction and solvent extraction.

2.3.5.2.1.1 Static headspace analysis

Direct sampling of headspace compounds involves injection of the vapour in

equilibrium with the sample held in a closed vial. This method has been used due to its

rapidity, simplicity, little sample preparation and reflection of the impression sensed by

the nose (Chang et al. 1977, Reineccius 1985, Josephson et al. 1987). However, the

drawback of this method is lack of sensitivity, since only the most abundant volatiles in

the food are detected (Reineccius and Anandaraman 1984, Min and Kim 1985). When

the sample size is increased to provide a larger sample of headspace gas, band

broadening and poor resolution is noted (Jennings, Wohleb and Lewis 1972). Therefore,

the problems encountered in this technique have been overcome by headspace

concentration techniques.

2.3.5.2.1.2 Dynamic headspace analysis

In headspace concentration techniques, the equilibrium headspace vapours of the

food are purged with inert gas, such as helium or nitrogen, by passing a stream of the gas

71

through the condensed material or swept/passed over the material. The volatile

compounds in the gaseous effluent are then trapped cryogenically or adsorbed in a

suitable trapping medium, such as activated carbon or synthetic porous polymers (Drozd

and Novak 1979, Reineccius 1994). With this technique, compound concentrations can

be increased 10-50 fold compared to that in the static headspace technique (Cronin

1982). Since there is a large contact area between the sample and the flowing gas during

purging, high recovery is commonly achieved (Przybylski and Eskin 1995). This method

has become the most popular method used in the isolation of flavour compounds from

foods (Reineccius 1989).

Although a cryogenic trapping procedure is a suitable method to analyse

compounds that are highly volatile, the major problem encountered in this method is the

large amount of water trapped concurrently, resulting in difficulty for subsequent

analysis. An additional step is necessary to extract the flavour constituents from the

water. The use of solvent, however, gives rise to concern about solvent impurities and

the opportunity of sample changes during flavour isolation, with the increasing step of

analysis (Reineccius 1994).

To avoid the problem of water condensation, trapping of flavour volatile

compounds can be effected by adsorbent materials, which have minimal water affinity

In this technique, volatile compounds are passed through a tube containing a small

amount of adsorbent material, where the volatiles are trapped and the adsorbed volatiles

are subsequently recovered either by heat desorption or solvent extraction. Some

adsorbents commonly used for trapping are activated carbon and synthetic porous

polymers such as Porapak Q (ethylvinylbenzene-divinylbenzene), Porapak P (styrene-

ethylvinylbenzene-divinylbenzene), Chromosorb (styrene-divinylbenzene) and Tenax GC

(poly-2,6-diphenyl-p-phenylene oxide) (Reineccius and Anandaraman 1984).

72

Activated carbon has the highest adsorptive capacity among the adsorbents

commonly employed. However, the formation of artifacts on carbon traps was noted

during thermal desorption and only high quality carbon produced excellent results when

desorption was effected by solvent extraction (Roeraade and Enzell 1972, Grob 1973).

The efficiency of adsorption of flavour compounds of many porous polymers has

been evaluated (Boyko, Morgan, and Libbey 1978, Withycombe, Mookheijee and

Hruza 1978, Dressier 1979, Sakaki et al. 1984). Although the capacity of Porapak Q

is relatively higher than that of Tenax GC, the reduced quantity of low molecular weight

alcohols retained by Porapak Q was noted (Jennings and Filsoof 1977). Moreover,

artifact peaks were also produced when Porapak Q was heated at 170°C and above

(Krumperman 1972).

Tenax GC, despite its lower adsorptive capacity due to a smaller surface area, has

been shown to yield excellent recovery of compounds. Its relatively high thermal stability

has also made it more suitable for high boiling point compounds than other synthetic

polymers, while the retention of water is very small, allowing insignificant water sorption

(Drozd and Novak 1979, Reineccius and Anandaraman 1984, Reineccius 1994).

Therefore, Tenax GC has become the adsorbent of choice in most studies.

Withycombe et al. (1978) found that Tenax GC gave the more organoleptically

characteristic isolate by comparison with either Chromosorb or Porapak Q. This was

supported by Buckholz et al. (1980) who noted that Tenax GC was useful for the

adsorption of headspace volatiles that represent a GC profile similar to that perceived by

the human nose in the original product. The use of Tenax GC to isolate volatile

compounds followed by solvent extraction to recover them and concentration under

nitrogen gas has been documented (Josephson et al. 1983, Olafsdottir, Steinke and

Lindsay 1985, Josephson et al. 1987), while thermal desorption of volatile components

73

trapped by Tenax GC has been reported by many workers (Sakaki et al. 1984, Vejaphan,

Hsieh and Williams 1988, Hsieh et al 1989, Matiella and Hsieh 1990).

2.3.5.2.1.3 Distillation/extraction

Isolation of aroma/flavour by distillation is also a popular technique instead of

headspace concentration (Reineccius 1989). In this technique, steam enters the vessel

containing the macerated sample which is agitated continuously while distilling off some

portion of water. Volatile compounds are then condensed in a cold trap chilled by ice

water or dry ice/acetone. Since the volatile compounds isolated are in very dilute

aqueous solution, extraction using an organic solvent is necessary to recover the volatile

compounds, followed by evaporation of the solvent to concentrate them. This is

commonly carried out by vacuum concentration, or purging with a gentle stream of

nitrogen. The addition of desiccant such as anhydrous MgSC>4 or Na2SC>4 has been

shown to be effective in the removal of any residual water from the system (Reineccius

and Anandaraman 1984). The solvent extraction of the distillate, as the additional step

that must be incorporated into the techniques, is the major disadvantage of this method

(Reineccius 1994).

Simultaneous steam distillation-extraction (SDE) originally proposed by Likens

and Nickerson (1964) has become a common method for the isolation of flavour

compounds from foods by distillation techniques (Cronin 1982, Reineccius 1994). The

major advantage of SDE over other distillation methods is that distillation and extraction

are carried out in a single operation. Various modifications of the original apparatus

have been developed and used extensively to isolate flavour volatiles from foods (Schultz

et al. 1977, Tanchotikul and Hsieh 1991, Cha and Cadwallader 1995).

The potential for artifact production in steam distillation was noted as possibly

74

due to the addition of antifoam agent, contaminated water supplies, solvent impurities

and thermally induced chemical changes (Reineccius 1994). However, when adequate

care is taken artifact generation can be minimised.

Solvent extraction directly on the food may also be used to isolate flavour

compounds from foods. The possibility of solvent to extract fat in this method is the

major concern. The extract containing flavour components and fats may be further

treated to obtain a flavour isolate. This increases the possibility of artifact production.

Therefore, this method is only suitable for flavour extractions from fat-free foods

(Reineccius 1994).

2.3.5.2.2 Separation of aroma components

When flavour components have been isolated, the next step is separation of the

components into individual compounds. Gas chromatography has become a very

important technique in flavour analysis due to its high sensitivity and resolution power.

The crucial components in GC analysis are the column, injector/inlet system and

detector.

2.3.5.2.2.1 Column

The development of flavour analysis significantly correlates to the development

of gas chromatographic column technology. The packed column is the oldest column

type used for separation. Although it can accept quite high amounts of sample without

overload, this column is low in sensitivity and resolution, resulting in unsatisfactory

separation for complex mixtures of flavour compounds.

The capillary column is characterised by a very small diameter with great length.

The liquid phase is distributed as a thin film on the wall of an open tube which results in

75

high permeability of the column, hence the analysis can be carried out in a shorter time

than that by a packed column. Moreover, there is no obstruction in the centre of the

tube, therefore much greater length can be applied in this column without excessive

pressure drop, resulting in much better separation (Dawes 1993, Grant 1996). For that

reason, capillary columns have become widely used for flavour analysis. Variations of

capillary column construction include porous layer open tubular (PLOT), support coated

open tubular (SCOT) and wall coated open tubular (WCOT) columns. Porous layer

open tubular columns have a relatively thick layer of particles of a solid adsorbent

material coated on to the capillary tube wall. The capacity of this column is relatively

larger than that of WCOT. This type of column has been used to analyse flavour

compounds of different types of food and analysis of the flavour from roasted peanuts

produced 142 compounds (Walradt et al. 1971). However, the SCOT column cannot be

easily repacked. The advanced technology of the WCOT column has made this type of

column more popular in the analysis of flavour compounds from foods and beverages.

2.3.5.2.2.2 Inlet system

According to Grant (1996), the major requirement of sample introduction is

introduction of the required quantity of sample that is representative of the original

without either overloading or lowering the performance of the column.

There are several techniques for sample introduction that are used. Recent

techniques that have been developed that are relevant to flavour analysis are splitless and

on-column injection (Reineccius 1985).

In splitless injection, nearly all of the sample and the solvent are applied to the

column with the split line closed. Then, the small amount of residual solvent and sample

that remains in the injection port is vented by opening the split line. The sample that has

76

been applied to the column is re-focussed near the head of the column by the solvent

effect technique, with simultaneous separation of the solvent from the sample. The

chromatogram is then run according to an appropriate temperature program. This

injection is suitable for trace level determination in dilute liquid samples (Tipler 1993).

In on-column injection, the liquid sample is applied directly to the column or to

the retention gap connected to the column. This is commonly used for the analysis of

either thermally labile or very high boiling compounds. Grant (1996) noted that on-

column injection might be the best method of sample introduction for capillary columns

regarding quantitation and trace analysis.

Another sample introduction technique developed for flavour analysis is that

using concentrating devices for sample injection. In these devices, the organic

components are trapped onto an adsorbent and the sample is subsequently thermally

desorbed on to the head of the GC column by reverse flushing with carrier gas (Kitson,

Larsen and McEwen 1996). Rapid increase of the temperature of the trapping device

can be performed for desorption after the volatiles are trapped by a simple cool stage.

This technique is mainly used to determine trace levels of organic pollutants in water,

however, recently this technique has been developed in a variety of applications for oils

and foods containing oil

2.3.5.2.2.3 Detector

There are a number of detectors that have been developed and those most

commonly used in GC are thermal conductivity detector (TCD), flame ionisation

detector (FID) and electron capture detector (ECD) and mass spectrometry (MS).

The first detector used for GC with universal response to all compounds was the

TCD. This is relatively insensitive, however, it can still be used to detect compounds

77

such as C02 that do not respond on more sensitive detectors.

The FID is now the most widely used detector in GC due to its high sensitivity,

stability and incomparable linearity (Flanagan 1993). This detector responds to a large

number of compound types, especially all organic molecules containing a C-H bond. The

response mechanism of this detector depends on mixing the carrier gas stream after

eluting from the column with sufficient hydrogen to produce a combustible mixture. In

an excess of clean air, the effluent mixture is burnt. When compounds from the column

enter the flame, they undergo combustion. During this combustion process, a very small

proportion of the carbon atoms undergoes ionisation and these ions are collected by an

electrode, that is polarised with respect to the jet. The resulting electrical current is

amplified to provide the chromatographic signal (Tipler 1993). The response of this

detector is dependent upon the number of C-H bond and not the number of molecules.

Compounds that do not contain carbon do not produce an FID response and the

presence of heteroatoms such as nitrogen, oxygen and sulphur tends to reduce the

detector response. Lack of FID response to water is particularly useful for the analysis

of aqueous samples.

The ECD is another important detector. This is classified as a selective detector

since it is extremely sensitive to halogenated hydrocarbons and certain other classes of

organic compounds such as conjugated carbonyls and nitrates, which have the ability to

accept a negative charge. This is a popular detector for the measurement of halogen

compounds particularly in environmental applications involving insecticide and herbicide

residues (Flanagan 1993).

The MS is a very important detection system used in GC. This selective detector

has the ability to separate gaseous positive ions according to their mass to charge ratio

(m/z). This is discussed in more detail in Section 2.3.5.2.41.

78

Other selective detectors include the nitrogen phosphorus detector (NPD) that

has selective response to phosphorus and nitrogen compounds. The flame photometric

detector (FPD) selectively detects phosphorus and sulphur-containing compounds and

the selectivity of the photo-ionisation detector (PLD) is based on the energy of the

emitted radiation.

2.3.5.2.3 Odour characterisation of individual flavour compounds

Although GC provides high sensitivity and resolving power, the usual GC

detector provides very little information for the flavour chemist since it cannot indicate

the flavour significance of each individual components that is separated. Hence it needs

the human nose to detect the flavour characteristic of those separated compounds. This

can be accomplished either by a non-destructive detector or an effluent splitter.

In an effluent splitter, GC effluent is split with the larger portion going to the

nose via an olfactory port and the other stream to the GC detector, therefore, a trained

person can smell the GC effluent and provide an idea of the sensory properties of

individual components of the sample, while the chromatogram is running.

The TCD is commonly used for odour profiling when a non-destructive detector

is used Since the entire sample is available for GC detection and smelling, this TCD is a

better choice than an effluent splitter, however, because of the lower sensitivity of this

detector compared with other GC detectors, an effluent splitter using FID is preferred.

One limitation of this technique is that it is a subjective method that might give

rise to inconsistent results. During a long analysis time odour fatigue might take place

leading to incorrect odour descriptions. However, by splitting the volatile compounds

into the GC detector and at the olfactory outlet, individual eluting compounds can be

characterised so that interesting compounds can be identified.

79

2.3.5.2.4 Identification of individual volatile compounds

When the flavour compounds have been isolated and separated, the last task in

flavour analysis is characterisation or identification. There are several methods for

identification of unknowns, however, since only small amounts (ng-qg) of compounds

from a large quantity of foods are available for identification, the suitable instrument for

identifications include GC, MS, infrared (IR) and nuclear magnetic resonance (NMR).

Although GC has an excellent capability to separate mixtures into individual, or

mostly individual components, it has only limited capability for identification. Thus,

ancillary techniques, such as homologue calculation, standards, Kovats indices,

substraction column are commonly used in helping GC for identification. Recently, GC

coupled with a second type of instrument such as IR and MS become popular instrument

for identification.

For identification purposes, IR spectra have presented certain potentialities to

provide information on functional groups present in the molecule. This is useful when the

analysis is effected using packed columns and concerns only major components,

however, this technique have become less popular when identification of trace

components in the flavour isolate is desirable (Reineccius 1985). The large sample size

required to obtain a useful spectrum also limits the use of this technique in flavour

analysis. The use of NMR spectroscopy to analyse flavour compounds is also limited due

to the large amount of sample required for analysis (Perkin 1975).

Mass spectrometry is another crucial instrument in conjunction with GC (GC-

MS) in flavour analysis. This instrument is generally used to determine the identity of

unknown compounds or as a mass selective detector (Reineccius and Anandaraman

1984). In comparison to IR and NMR, MS requires much less sample for analysis

(Perkin 1975).

80

2.3.5.2.4.1 Mass spectrometry

The mass spectrometer is a device that provides electrical detection and has the

ability to separate gaseous positive ions according to their mass to charge ratio m/z and

measure the abundance of each ion species (Kitson et al. 1996). Molecules in the gas

phase are ionised by electron bombardment or by other chemical ionisation methods.

These fragment ions are then accelerated with a high electrical potential into an analyser

which separates the ions according to their m/z. The separated ions then are detected

and their intensity measured.

The principal constituents of a mass spectrometer are an ion source, where ions

are formed from the sample, an analyser which separates the ions according to their m/z

values, a detector which gives intensity of the ion current for each species, electronic

power supply and control of the various units and the vacuum pumping system

(Constantin and Schnell 1990).

There are several techniques used to convert the initially neutral sample into an

ionised species in the gas phase, but the most widely used for GC-MS are electron

impact ionisation (El) and chemical ionisation (Cl) (Evershed 1993, Kitson et al. 1996).

In El, the effluent from the GC enters a partially enclosed ion source. Electrons

boiled from a hot wire or ribbon, are accelerated by a potential difference typically of

70eV. When these electrons pass near neutral molecules, they may impart sufficient

energy to remove the outer shell electron, producing an additional free electron and

positive (molecular) ions. The high energy imparted can also cause the molecular ion to

break apart into further neutral atoms and fragment ions.

In the Cl method, the reagent gas enters to the ion source at a pressure of about

1 torr (mm Hg). When the sample having a partial pressure of about 0.01 torr or less is

81

admitted and the electron beam is passed through the mixture, the reagent gas ions are

preferentially formed and undergo ion-molecule reactions with the sample (Constantin

and Schnell 1990). With this method, less fragmentation and more abundant molecular

ions are usually produced, because less energy is transferred during ionisation

(Gudzinowicz, Gudzinowicz and Martin 1976).

Mass analysers, commonly used in GC-MS instruments, are quadrupole mass

filters and magnetic-sector analysers (Evershed 1993, Gudzinowicz et al. 1976, Kitson

et al. 1996).

In the magnetic sector instrument, gas phase ions produced in the ion source are

accelerated by thermal energy through a potential gradient. These ions travel through a

vacuum chamber into a magnetic field at a sufficiently low pressure so that collisions

with neutral gas molecules are uncommon and all of the ions entering the magnetic field

have approximately the same kinetic energy. Before striking a detector, the ions pass

through a fixed slit, and sweeping the magnetic field from high to low field allows ions of

progressively lower mass to pass through the slit and strike the detector.

The quadrupole mass filter is based on an alternating electric field applied to four

parallel rods of hyperbolic, elliptical or circular cross-section held in a square

(quadrangular) array . From the ion source, ions enter the quadrupole and are sorted by

imposing radio frequency (rf) and dc voltage on diagonally opposed rods. When the rf

and dc voltages are swept in a fixed ratio, ions with higher masses follow a stable path to

the detector and at a certain field strength, only ions in a narrow m/z range reach the

detector while others are deflected into the rods.

For GC-MS systems, the quadrupole analyser is popular due to its capability for

fast scanning, high transmission rate resulting in high sensitivity and independence from

the energy distribution of the ion beam (Evershed 1993, Gudzinowicz et al. 1976).

82

Ion detection can be performed by 2 methods, i.e. photographic and electrical

detection (Constantin and Schnell 1990). Photographic detection acts simultaneously as

collector, converter and recorder. For a particular instrument, this detector gives the

highest possible resolving power and can give simultaneous recording of the whole

spectrum. Compared to electrical detection, this method is much less sensitive and the

response is not a linear function of the energy or number of ions. Therefore this is not

suitable for measurement of isotopic abundance and fragmentation studies.

There are 2 main types of collector system used in electrical detection, i.e. the

Faraday cup and the electron multiplier. The latter is commonly used for ion detection in

GC-MS. The basic principle of electron multipliers is the emission of secondary

electrons when an energy particle strikes a suitable surface. The secondary electrons can

be multiplied by being made to strike a second such surface and so on. The cascade of

electrons will increase when the process is repeated and at the end of the cascade

sufficient signal is provided to be detected.

2.3.5.2.4.2 Gas chromatography-mass spectrometry (GC-MS) coupling

The combination GC-MS has become a powerful tool in the analysis of a

mixture and the structural determination of unknown compounds. As an excellent

separator, the ability of GC to identify is limited and does not produce any structural

information. On the other hand, although MS is an excellent tool for structural

characterisation, it has no separation capability.

The main obstacle in connecting these two instruments is that MS operates at a

high vacuum, i.e. 10"5-10"8 torr, while the GC effluent is at atmospheric pressure, i.e. 760

torr, with the flow rate of 1-2 mL/min. Therefore reducing the pressure from 760 torr at

the column outlet to 10"5-1 O'6 torr in the ion source housing is necessary. This can be

83

accomplished by using a high resolution GC-MS interface such as a jet separator (Perkin

1989) which consists of two capillary tubes connected with a small space (approx. 1

mm). By using a rotary pump, a vacuum is created between the tubes and GC effluent

passes through one capillary tube into the vacuum region. Molecules that continue in the

same direction will enter the second capillary tube and will be directed to the ion source

(Kitson e/a/. 1996).

When capillary columns are used in the GC, the direct connection without the

need for an interface is common practice (Grant 1996, Kitson et al. 1996). To maintain

the source pressures required for mass spectrometry, a large capacity diffusion pump is

applied (Perkin 1989) and, under normal operating conditions, the mass spectrometer

pumping system can handle the whole of the effluent from the column (Kitson et

al. 1996). To prevent sample condensation, it is necessary to heat the capillary column

between the GC and the MS ion source above the boiling point of the highest boiling

point compound of the sample (Kitson et al. 1996). According to Grant (1996), the only

concern with this method is that column performance can be adversely affected due to

the high vacuum at the end of the column. This can be avoided by placing a restrictor

between the end of the column and the ion source so that most pressure drops below

atmospheric pressure take place in the restrictor

84

3 MATERIALS AND METHODS

3.1 Materials

All fish used in the experiments were Australian sardines bought from the Sydney

Fish Market. The forms and conditions of the fish were different for each experiment, as

indicated.

3.2 Equipment

The equipment used in the experiments are listed in Table 3.1.

Table 3.1. Equipment used in the experiments

Equipment Manufacturer and type

Sampling tube SGE Sci. Pty Ltd, AustraliaAnalytical balanceAw meter Vaisala Humidity & Temperature Indicator HMI 31Capillary column DB-23, 60m length, 0.25 i.d., J&W Sci. Inc.USAConway dishesElectrical tunnel dryerFluorescence spectrophotometer Hitachi F 1000Food protein analyser Leco FP 448, leco Aust. Pty Ltd.Forced air ovenGLC Hewlett-Packard, USA, Model 5890Headspace injector SGE Sci. Pty Ltd. AustraliaHumidity chamberIntegrator Hewlett-Packard, USA, 3392AMetallised biaxially oriented

polypropylene packing sheetMicro-biuretMini column Silica-cartridge 2000mg/mlMuffle furnaceOven drying (air oven)pH meter Orion Research Inc., Digital Ionalyzer, Model 601ASilica gel plates Whatman/Merck, LK 6, 20cm x 20 cmSpectrophotometer Shimadzu, UV-120-02Table top centrifuge EppendorfVacuum rotary evaporator Rotavapor R110, BuchiVacuum sealerWaring blender Waring

85

3.3 Chemicals and solvents

Table 3.2 Chemicals and solvents used in this experiment

Chemicals/solvents Brand and specification

Solvents’.

Acetic acid, glacial BDH, AR

Chloroform Ajax Chemicals, AR (redistilled)

Dichloromethane Ajax Chemicals, AR

Diethyl ether Ajax Chemicals, AR (redistilled)

Hexane Ajax Chemicals, AR (redistilled)

Isooctane Ajax Chemicals, AR

Isopropanol Mallinckrodt, AR (redistilled)

Methanol Mallinckrodt, AR (redistilled)

Chemicals’.

Boric acid E. Merck, AR

Boron trifluoride Supelco

Bovine serum albumin Sigma

Bromocresol green BDH, GR

Copper sulphate anhydrous E. Merck, GR,

Cupric acetate E. Marck, AR

Folin-Ciocalteu phenol Sigma

Formaldehyde AR Ajax Chemical

Glycerol Ajax Chemical

2-Heptanone Sigma

Hydrochloric acid E. Merck, GR

Hydrogen gas -

Lithium chloride E. Merck, GR

Magnesium chloride E. Merck, GR

Magnesium nitrate E. Merck, GR

Magnesium perchlorate E. Merck

Methyl red BDH

cont

86

Table 3.2 Chemicals (continued)

Nitrogen gas

Orthophosphoric acid

Potassium bicarbonate

Potassium chloride

Potassium chromate

Potassium sulphate

Quinine sulphate Silver nitrate

Sodium bicarbonate

Sodium chloride

Sodium chloride

Sodium dodecyl sulphate

Sodium hydroxide

Sodium tartrate TenaxR TA

Trichloroacetic acid (TCA)

Tricosanoic acid

Undecane

Mallinckrodt

E. Merck, AR

E.Merck

May & Baker Ltd., AR

E.Merck

Sigma

BDH, AR

BDH, AR

Pacific salt, kiln dried fine salt

Jaegar Chemicals

Bio-Rad

E. Merck, GR

E. MarckAlltech (Australia) Pty. Ltd. , 60/80 mesh

E.Merck, GR

Sigma

Sigma, 99%

3.4 Selection of Extraction Method for Fish Lipids

3.4.1 Raw materials

Raw material used in this experiment was either fresh or frozen sardines bought

from the Sydney Fish Market, and road transported in ice to the laboratory of The

Department of Food Science and Technology, The University of New South Wales,

Kensington, NSW. Upon arrival at the laboratory the fish were beheaded, gutted and kept

frozen (-25°C) until used. Frozen samples from the market were kept at -25°C. Since

evaluation of the extraction method was conducted over several months, the condition of

the raw materials and the time of purchase differed as stated for each experiment.

87

Solvents used in this experiment were analytical grade and redistilled before use.

3.4.2 Methods

3.4.2.1 Extraction methods

3.4.2.1.1 Bligh and Dyer (1959) method

Fish was blended in a Waring blender with chloroform:methanol (1:1:2 w/v/v) for 2

min. Chloroform (1 part) was added to the mixture, blended for 30 sec, followed by the

addition of 1 part of distilled water and blended for another 30 sec. The homogenate was

suction filtered through a Whatman No 1 filter paper on a Buchner funnel and the residue

reblended with 1 part of chloroform, filtered through the Buchner funnel used in the first

filtration, and the residue was rinsed with 0:5 part of chloroform. This filtrate was mixed

with the first filtrate, transferred to a separating funnel, and left to stand until the layers

separated completely. The chloroform phase consisting of total lipid in the lower layer was

vacuum rotary evaporated at 40°C. The residual fat was redissolved in a small amount of

chloroform, filtered and the filtrate re-evaporated in a similar manner.

3.4.2.1.2 Hara and Radin (1978) method

One part of fish was blended with 5-50 parts (depending on the ratio of tissue to

solvent) of a mixture of hexane and isopropanol (3:2 v/v) for 2 min, and the homogenate

filtered through a Whatman No 1 filter paper on a Buchner funnel. The residue was washed

three times with the same solvent (2 parts of washing solvent for 18 parts of the original

solvents), by re-suspending and soaking the residue for 2 min before filtering. The filtrate

containing total lipid was vacuum rotary evaporated at 40°C. The residual fat was

redissolved in a small amount of chloroform, filtered and the filtrate was re-evaporated in a

similar manner.

88

3.4.2.2 Preparatory method

Before comparing the Bligh and Dyer (1959) method with that of the Hara and

Radin (1978) method, the effect of raw material preparation before extraction on the

amount of lipid recovery as well as optimisation of the Hara and Radin (1978) method

were carried out.

3.4.2.2.1 Effect of raw material preparation

For lipid extraction from individual fish, the frozen fish was chopped individually,

and 20 g was taken from an individual chopped fish for each extraction method (Bligh and

Dyer 1959, Hara and Radin 1978 methods as indicated in Sections 3.4.2.1.1 and 3.4.2.1.2

respectively) and for each replication. The proportion of tissue to solvent in the Hara and

Radin (1978) method was 1:5 w/v.

For lipid extraction from blended fish, 5 frozen fish were blended, pooled and 20 g

taken for each extraction method (Bligh and Dyer 1959, Hara and Radin 1978 methods)

and replication. The extraction processes using both Bligh and Dyer (1959) and Hara and

Radin (1978) methods were similar to those indicated in Section 3.4.2.1.1 and 3.4.2.1.2,

respectively.

Lipid recovery both from individual fish and blended fish were determined

gravimetrically.

3.4.2.2.2 Optimisation of Hara and Radin (1978) method

Different proportions of tissue to solvent (1:5 and 1:10 w/v) were evaluated

regarding the amount of extracted lipid, and further evaluation of protein and polar

compounds were carried out for other tissue : solvent proportions (1:10, 1:20, 1:30, 1:40,

1:50 w/v). The process of extraction was similar to the previous method for the Hara and

Radin (1978) method as indicated in Section 3.4.2.1.2, except for the ratio of tissue to

89

solvent.

A study of the effect of washing in extraction with different ratio of tissue to solvent

(1:10, 1:20, 1:30, 1:40, 1:50 w/v) was also carried out. Parameters assessed were extracted

lipid amount, protein contaminant and the level of polar compounds.

The process of extraction without washing was similar to the previous method of

Hara and Radin (1978) as indicated in Section 3.4.2.1.2 except for the proportion of

tissue to solvent.

In the extraction method applying washing, before evaporation, the filtrate

recovered was placed in a separating funnel, aqueous sodium sulphate (0.6 volume) added,

and shaken vigorously for 1 min and left to stand until the layer separated completely. The

upper, hexane-rich layer containing the lipid was separated and evaporated by vacuum

rotary evaporator at 40°C.

3.4.2.2.3 Evaluation of Bligh and Dyer (1959) and Hara and Radin (1978) methods

The optimum proportion of tissue to solvent in the Hara and Radin (1978) method

(1:20 w/v) was compared to the Bligh and Dyer (1959) method regarding the quantitative

recovery, qualitative purity and the stability of lipid recovered. Protein contamination and

polar compounds were determined and the lipid stability was evaluated by observing

changes in fluorescence and lipid weight during accelerated oxidation at 60°C.

3.4.2.3 Analytical methods

3.4.2.3.1 Moisture content

Moisture content was analysed according to Aitken, Lees and Smith (1984). Two

to five grams of sample were weighed accurately in a dried and tared crucible dish and

placed in an air oven (101 °C; 24 h). The dishes containing sample were transfered to a

90

desiccator, allowed to cool and weighed. Weight loss after drying was considered as the

moisture content of the sample expressed as % (w/w).

3.4.2.3.2 Lipid content

Lipid content was measured gravimetrically. A volume of solvent containing lipid

extracted either by the Bligh and Dyer (1959) or the Hara and Radin (1978) methods was

evaporated to dryness in a tared flask using a vacuum rotary evaporator at 40°C and the

weight of lipid residue determined. The lipid content (% w/w of fish sample) was calculated

based on the original volume of solvent.

3.4.2.3.3 Soluble protein

Protein contamination in lipid extracts (in chloroform) was determined by the

method of Lowry et al. (1951) modified by Markwell et a/. (1978). A volume (100-1000

pi) of chloroform containing a known weight of lipid was mixed with 3 mL reagent C in a

test tube and incubated for 20 min at ambient temperature. Reagent C was made from a

mixture of 100 parts reagent A (2.0% sodium bicarbonate, 0.4% sodium hydroxide, 0.16%

sodium tartrate and 1% sodium dodecyl sulphate/SDS) and 1 part reagent B (4%

CuS04.5H20). After incubation, 0.3 mL of Folin-Ciocalteu (phosphomolybdic-

phosphotungsic acid) reagent (1:2 v/v) was added, mixed vigorously and incubated for 45

min at ambient temperature, then 3.0 mL dichloromethane was added, mixed and

centrifuged at 5000g- for 5 min using a table top centrifuge. The absorbance of the top

layer was read at 660 pm and protein content was determined using bovine serum albumin

as the standard.

3.4.2.3.4 Polar compounds

The polar compounds were determined by the method of Junaeda and Rocquelin

91

(1985). Fifty to ninety mg oil was dissolved in 2 mL chloroform and passed through a

silica Sep-Pak cartridge. About 30 mL chloroform was used to elute non-polar lipid (NPL),

followed by 20 mL mixture of chloroform and methanol (49:1 v/v) to elute all

monoacylglycerols. The fraction containing polar compounds was eluted with

approximately 25 mL of methanol, evaporated and weighed.

The purity of the fractions was verified by thin layer chromatography (TLC) on

silica gel plates (Whatman or Merck). The developing solvent system was a mixture of

hexane-diethyl ether-methanol-acetic acid (90:20:5:2 v/v). The plate was sprayed with 3%

cupric acetate in 8% orthophosphoric acid and charred for 30 min at 120°C.

3.4.2.3.5 Fluorescent products

The analysis of fluorescent components was carried out according to the methods of

Fletcher et al. (1973) modified by Lubis (1989). About 60-100 mg lipid was weighed

accurately and dissolved in 10 mL chloroform, then exposed to a source of UV light to

destroy any retinol present. The fluorescence was measured using a

spectrophotofluorometer with excitation of 370 nm and emission of 460 nm using 1 fig

quinine sulphate in 1 mL 0.1N sulphuric acid as a standard for fluorescence intensity.

3.4.2.3.6 Weight gain

Weight gain measurement was carried out by containing approximately 2 g lipid in

Petri dishes (ID 50 mm) and placing them in a forced air oven at 60°C. The samples were

periodically taken out, placed in a desiccator and weighed The changes of weight were

monitored daily.

92

3.5 Simulation of Fish Spoilage

3.5.1 Raw materials

The raw material used in this experiment was fresh sardine (Sardinops

neopilchardus) bought from Sydney Fresh market, Pyrmont, in June-July 1995,

transported in ice to the laboratory at UNSW and used directly.

3.5.2 Methods

3.5.2.1 Preparatory methods

Fresh sardines were placed in a humidity chamber, set at RH 70-80% and 28-30°C

(simulation of the Indonesian climate) for 14 h and samples were taken after the first 4 h

and every 2h afterwards. Samples were evaluated for sensory assessment, moisture

content, pH and total volatile base nitrogen (TVB-N). The experiment was replicated 3

times.

3.5.2.2 Analytical methods

3.5.2.2.1 Sensory assessment

Sensory evaluation was carried out on the ungutted fish using the Tasmanian Food

Research Unit (TFRU) scheme in which scoring is based on awarding demerit points for

loss of quality as attached in Appendix (Table 7.2.6). Many attributes commonly used

to judge fish are assessed. Each attribute consists of a score from 0 to 3, where 0

represents good quality and higher scores indicates progressively poorer quality. When the

scores of all attributes are summed, zero or near zero indicates that fish is very fresh, while

39 represents the worst quality of ungutted fish (Branch and Vail 1985).

93

3.5.2.2.2 Moisture content

Moisture content was analysed as described in Section 3.4.2.3.1.

3.5.2.2.3 pH

pH was measured using a pH meter which had been calibrated with standard pH

buffers 4 and 7. Fifteen grams of sample were mixed with 15 mL of distilled water and the

pH of the solution was measured (AO AC 1990).

3.5.2.2.4 Total volatile base nitrogen (TVB-N)

Total volatile base nitrogen was analysed by Conway’s microdiffusion method

(Siang and Kim 1992). Four grams of sample were mixed with 16 mL of 4%

trichloroacetic acid (TCA), blended well, allowed to stand for 10 min, filtered through

Whatman fiter paper No. 1, and the filtrate analysed for the volatile base nitrogen as

follows.

Into the inner ring of the Conway unit, 1 mL of 1% boric solution containing

bromocresol green and methyl red indicators was pipetted. One mL of sample filtrate was

pipetted into the outer ring, then 1 mL of saturated K2C03 added and the Conway unit

covered immediately. The outer ring solution was mixed gently and the Conway dishes

were incubated at 37°C for 60 min. After incubation, the inner ring solution was titrated

against 0.0IN HC1 using a micro-burette until the green colour turned pink. If the analysis

could not be performed immediately after filtration, the filtrate was frozen (-20°C).

3.6 Lipid Characterisation of Raw Materials and Their Dried Salted Products

3.6.1 Raw materials

Australian fresh sardines were purchased from the Sydney Fish Market, Pyrmont, in

October 1995 and road transported in ice to UNSW.

94

3.6.2 Methods

3.6.2.1 Preparatory method

Upon arrival at the laboratory, fish were washed, wrapped individually in metallised

biaxially oriented polypropylene (BOPP), vacuum packed and kept frozen

(- 25°C) until used within one month.

3.6.2.2 Simulation of fish spoilage

Fish were thawed in a chilling room (5° to 7°C), placed in a humidity chamber as

described in Section 3.5.2.1 by which moderately spoiled or spoiled raw materials were

obtained after 6h and lOh incubation, respectively. Fresh and deteriorated fish were

sensorily evaluated and analysed for moisture content, TVB-N and trimethylamine (TMA-

N) as described in Sections 3.5.2.2.1, 3.4.2.3.1, 3.5.2.2.4 and 3.6.2.5.1, respectively.

3.6.2.3 Processing of dried salted fish

Raw materials (fresh, moderately spoiled and spoiled raw material) were salted

separately in saturated brine (fish to brine ratio was 1:2 w/v) for 12 h with addition of salt

crystals to maintain saturation. Upon completion the fish was washed with tap water and

then dried using a tunnel dryer at 40°C, RH 30% and 2 m/sec air velocity until the moisture

content of the sample was in the range of 44-46% (Lubis and Buckle 1990), which was

determined based upon the initial moisture content and weight changes. After drying, the

fish was air cooled, vacuum packed and kept frozen (-25°C) until analysed.

3.6.2.4 Lipid extraction

Oil was extracted from each type of raw material and its dried salted product using

the Bligh and Dyer method as described in Section 3.4.2.1.1. The solvent containing

lipids was pooled and then contained in small bottles (2mL each) which were flushed with

95

nitrogen, sealed and kept frozen (-25°C) until analysed. Lipids from raw materials and

dried salted fish were characterised for lipid content, polar and non-polar material,

fluorescence, polyene index and volatile compounds.

3.6.2.5 Analytical methods

3.6.2.5.1 Trimethylamine nitrogen (TMA-N)

Trimethylamine nitrogen was analysed using Conway’s microdiffusion method

(Siang and Kim 1992) as those for TVB-N except the addition of 1 mL of neutralised 10%

formaldehyde into the outer ring before the addition of saturated K2C03.

3.6.2.5.2 Lipid content

Lipid was extracted by the method described in Section 3.4.2.3.2.

3.6.2.5.3 Polar and non-polar compound

Polar and non-polar compounds were separated and determined by the method

described in Section 3.4.2.3.4.

3.6.2.5.4 Fluorescence

Fluorescence products were measured using the method described in Section

3.4.2.3.5.

3.6.2.5.5 Polyene Index

Polyene index as the ratio of (C20:5 + C22:6) and Cl6:0 (Lubis and Buckle 1990)

was calculated from the result of the fatty acid methyl esters (FAMEs) analysis. FAMEs

were determined according to AOCS (1990) Official Method Ce lb-89. To calculate the

absolute concentration of the individual fatty acids, tricosanoic acid (C23.0) as an internal

standard and theoretical relative response factors (TRF) from Craske and Bannon (1988)

were used.

96

A Hewlett-Packard GLC model 5890 (Hewlett-Packard, Avondale, PA, USA),

fitted with a flame ionisation detector (FID) and a Hewlett-Packard 3 3 92A integrator were

used to analyse the FAME. Capillary column of DB-23 (50% cyanopropyl/50%

methylpolysiloxane) with length of 60 m, i d. 0.25 mm and film thickness 0.25 pm (J & W

Scientific, Inc., Folsom, CA, USA) was used for separation.

FAME was injected manually into the GLC using the quantitative technique

developed by Craske (1995) with initial temperature of 120°C which was maintained for 0.5

min then increased to 170°C at a rate 40°/min, then further increased to 240°C at a rate of

1 °/min and maintained for 20 min.

The GLC was set as follows: injector temperature 350°C; detector temperature

250°C, gas pressure and flow rate: 120 kPa, 34 cm/sec (0.91 mL/min) for carrier gas (H2);

110 kPa, 30 mL/min for make up gas (N2), 110 kPa, 24 mL/min for fuel gas (H2) and 260

kPa, 330 mL/min for air pressure. The integrator parameters were: attenuation 3,

threshold 2, area rejection 1000, peak width 0.04 and chart speed 0.5 cm/min.

3.6.2.5.6 Headspace/volatile compounds

The volatile compounds were determined according to Hsieh et a'l. (1989) with

modification.

Approximately 50 mg oil was weight accurately and placed in the sampling

apparatus (a pear shape, 2 neck flask). The oil sample was prepurged without heating for 2

min, and then purged with nitrogen at a flow rate of 50 mL/min for 60 min at 65 ± 2°C, and

the volatiles were trapped in a sampling tube filled with 100 mg Tenax® TA. After 60 min

of purging and trapping, the nitrogen gas flow was directed through the Tenax® TA trap

for another 10 min. The trapped volatiles were transferred onto the GC column (carbowax

20M = polyethylene glycol of 20000 MW; J&W Sci., Inc., Folsom, CA) by inserting the

97

sampling tube in the headspace injector (SGE Scientific Pty Ltd, Australia) on the top of the

GC, and thermally desorbed at 185°C for 15 min with a stream of carrier gas. The volatile

compounds were then recondensed and focussed cryogenically onto a short section of

column by cooling the section with liquid carbon dioxide prior to chromatographic analysis.

The GC running program was as follows: initial temperature 30°C for 2 min, then

increased to 175°C at l°C/min and finally increased to 240°C at 20°C/min and maintained

for 30 min. The injector temperature was 185°C, detector temperature 250°C and the

pressure and flow rate of gases were 120 kPa, 34.0 cm/sec (1.0 mL/min) for carrier gas

(H2), 110 kPa, 24 mL/min for make up gas (N2), 110 kPa, 32 mL/min for fuel gas (H2) and

260 kPa, 300 mL/min for air pressure. The integrator parameters were: attenuation 1,

threshold 0, area rejection 1000, peak width 0.04 and chart speed 0.5 cm/min.

3.7 Characterisation of Dried Salted Fish Flavour

3.7.1 Raw materials

Raw materials used for this experiment were the same as those described in Section

3.6.1.

3.7.2 Methods

3.7.2.1 Preparatory methods

Preparatory methods used in this experiment (simulation of fish spoilage and

processing of dried salted fish) were the same as those described in Sections 3.6.2.2 and

3.6.2.3.

3.7.2.1.1 Volatile compounds

Dried salted fish were beheaded, eviscerated, blended, contained in BOPP pouches,

vacuum sealed and kept in the freezer until analysed.

98

3.7.2.1.2 Sample preparation for sensory evaluation

Dried salted fish were beheaded, eviscerated, cut into small pieces (1x1.5 cm) and

fried in coconut oil at 180°C for 40 sec. The coconut oil was always new for each batch of

frying; six beheaded and eviscerated fish that were cut into small pieces were fried in each

batch of frying.

3.7.2.2 Analytical methods

The analytical methods used in assessing fish spoilage were the same as those used

in Section 3.5.2.2.

The dried salted fish prepared from different raw material qualities were analysed for

TVB-N (Section 3.5.2.2.4 ), TMA-N (Section 3.6.2.5.1), moisture (Section 3.4.2.3.1),

lipid (Section 3.4.2.3.2), ash, salt, aw, volatile compounds and sensory characteristics.

3.7.2.2.1 Protein

Protein content of dried salted fish was measured based on the total nitrogen using a

Leco FP 448 Food Protein Analyser (Leco Aust. Pty Ltd.). The sample weights were

about 0.25 g. The conversion factor from nitrogen to protein was 6.25.

3.7.2.2.2 Ash

Ash content was measured according to Miwa and Ji (1992). Approx. 5g of sample

was weighed accurately in a tared crucible dish. The dish containing sample was dried

overnight in an oven at 105°C then ashed overnight in a furnace (550°C). The dish

containing dried sample was then placed in a furnace and heated overnight at 550°C. When

the colour of the ash had changed to white, the dish was taken out from the furnace, placed

in a desiccator for approximately 30 min and weighed. Ash content was expressed as % by

weight.

99

3.7.2.2.3 Salt

Salt content was determined according to FAO (1981). Approximately 2 g of

sample was accurately weighed and macerated in distilled water for about 2 min. The

extract was transferred quantitatively into a 250 mL volumetric flask and made up to

volume. Twenty-five mL aliquots were titrated against 0.1N silver nitrate using potassium

chromate as an indicator. The salt content was calculated as follows:

Titr.vol. sample - titr. vol. blank% salt = ---------------------------------------------x 5.844

Sample weight

3.7.2.2.4 Water activity (aw)

Water activities of samples were measured using an aw meter (Vaisala Humidity &

Temperature Indicator HMI 31). Before measuring the sample, the aw meter was calibrated

using RHs of standard saturated salt solutions, i.e. LiCL (11%), MgCl2.6H20 (33%),

Mg(N03)2.6H20 (55%), NaCl (75%), KC1 (85%) and K2S04(97%).

3.7.2.2.5 Flavour/volatile compounds

Flavour or volatile compounds of dried salted fish were determined using the

dynamic headspace sampling technique as indicated in Section 3.6.2.2.2 except that the

weight of sample was increased to approximately 1.5 g.

The GC program for separation in this experiment was also the same as that

described in Section 3.6.2.5.5 except for attenuation and area rejection which were set at 0.

3.7.2.2.6 Sensory assessment

The sensory evaluation was based on preference test (hedonic test) including

overall acceptability for raw and fried fish assessed by Indonesian panelists. Raw products

were assessed for appearance, colour, texture and aroma, while fried product was assessed

100

only for aroma and taste using the score sheet shown in Appendix (Table 7.4.9).

3.8 Statistical analysis

Experimental data were analysed by analysis of variance using a statistical computer

package (Statistica Rel. 5, StatSoft, Inc., USA).

101

4 RESULTS AND DISCUSSION

4.1 Selection of Extraction Method for Fish Lipids

Since fish lipids are very labile and are susceptible to oxidation, all solvents used in

this experiment were redistilled and stored under nitrogen to prevent oxidation of fish lipid

by peroxide residues in the solvents and to remove traces of non-volatile residues.

In order to ensure representative data, the effect of raw material preparation on the

lipid recovery in fish oil extraction by the methods of Bligh and Dyer (1959) and Hara and

Radin (1978) was investigated. Lipid extracted from individual fish showed variation in the

amount of lipid recovered, reflecting the heterogeneity of lipid content among individual

fish, while the amount of lipid extracted from pooled blended fish was not significantly

different among replications (Table 4.1).

Table 4.1 Lipid recovered (g/lOOg) from different forms of fish (individual & blended)

Replicate** Individual fish Blended fish

Bligh and Dyer

1 1.47 + 0.04 a 2.44 ±0.04 c2 1.16 ± 0.01 b 2.42 ±0.03 c'J 1.21 ±0.02 b 2.47 ±0.03 c

Hara and Radin*

1 1.58 + 0.01 k 2.18 ±0.03 m2 1.48 + 0.02 k 2.17 ± 0.01 m3 3.89 ±0.06 / 2.17 ±0.02 m

* = tissue:solvent = 1:2.5 (w/v)** = each replicate consists of 3 measurementsValues followed with the same letter/s within a column are not significantly different (p>0.05) No correlation among letters in different extraction methods

102

The heterogeneity of lipid content among individual fish might result from different

sexes and ages although visually the fish were uniform. Lipid content and fatty acid

composition are highly variable from fish to fish within species, due to biological differences

(Stansby 1982, Krzynowek et al. 1992, Sigurgisladottir and Palmadottir 1993, Love 1994).

Moreover, only a few fish were taken for analysis resulting in a high probability of variation

from fish to fish in the amount of lipid recovered. By blending all fish before extraction, the

variation encountered between replications was significantly reduced.

Before comparing the extraction method of Bligh and Dyer (1959) with that of

Hara and Radin (1978), optimisation of the latter method was carried out regarding the fish

to solvent ratio.

In the original method of Hara and Radin (1978), the ratio of tissue to solvent was

1:18 (w/v) with one phase of filtrate. When the proportion of fish to solvent was increased,

i.e. 1:2.5 and 1:5 (w/v), two layers were formed in the filtrate, in which the hexane layer

was only 50% and 70-80% from the total filtrate, respectively, and after the hexane layer

was separated and evaporated, the lipid recovery was low. It seems that, when the

proportion of fish to solvent was 1:2.5 and 1:5 (w/v), the lipid was not extracted

completely. According to Nelson (1991), in high fish to solvent ratio, the water content of

the solvent was higher and less organic phase was present, hence, less non-polar lipid was

extracted into the organic phase.

The single phase filtrate started to form when the ratio of tissue to solvent was

decreased to 1:10 (w/v), thus, with this ratio, the aqueous phase did not separate from the

organic phase, and consequently no aqueous phase was discarded. This was in accordance

with the original method of Hara and Radin (1978), which used a single phase solvent

system to extract lipid from tissue. Further trials to investigate the optimum ratio of fish to

solvent were carried out by increasing the proportion of solvent, i.e. fish to solvent ratios of

103

1:20, 1:30, 1:40 and 1:50).

Total lipid recovery extracted by the Hara and Radin (1978) method increased with

an increase of the proportion of solvent to tissue (Fig. 4.1). The lowest amount of lipid

recovered was produced by a tissue to solvent ratio of 1:10 (w/v), while the levels of

protein contaminant and polar lipid extracted by the same proportion were the highest.

With decreasing fish to solvent ratio, the lipid recovery increased and the level of impurities

decreased. This was as expected, since increasing the solvent to fish ratio will decrease the

water content of the solvent system, thus less water soluble impurities will be carried over

into the organic phase and more non-polar lipid will be extracted into the organic phase

(Dambergs 1959, Nelson 1991).

c£Tissue:solvent ratio (w/v)

Figure 4.1 Lipid yield, polar compounds and protein content of lipid extracted by the Hara and Radin (1978) method without washing

polar lipid (g/lOOg lipid): •. total lipid (g/lOOg fish): ♦. and protein (mg/g lipid): I

According to the original method, the hexane/isopropanol extract can be processed

without washing. However, in this study, when extracted lipid was redissolved in

104

chloroform, the solution was cloudy, even after being centrifuged, resulting in difficulty

during further analysis using a spectrophotometer. It was, therefore, necessary to include a

washing step in this method. Although contaminants in the lipids extracted by the method

without washing could be decreased by redissolving the lipids in chloroform and refiltering,

it might still contain non-lipid substances. Christie (1982) noted that separation of non-lipid

contaminants with such a method (redissolved in non-polar solvent and refiltered), was

rarely complete and non-lipid contaminants were still high (25-75%) and needed to be

removed by a reliable method (Nelson 1991).

When separate washing with sodium sulphate solution (6-7%) was applied to the

extracted lipid, the filtrate containing lipid was clear. As in the case of the same method

without the washing step, different proportions of tissue to solvent in this method resulted

in different amounts of lipid recovery. The lower the proportion of tissue to solvent, the

higher the lipid recovery.

The lipid recovered in this method, with the washing step, was 14-18% lower than

that obtained without washing (Fig. 4.2). It was suspected that more impurities were

carried over into the filtrate when extraction was performed without washing. This was

supported by the data that shows large differences in the level of polar compounds and

protein contaminants between extracts without and with washing (Figs. 4.3 and 4.4). The

washing step applied in this method decreased the protein contaminant effectively, i.e. 13 -

25 times lower than that without washing, and there was no significant difference (p>0.05)

in the level of protein contaminant obtained with different proportions of tissue to solvent.

Besides removing protein effectively, washing applied in the Hara and Radin (1978)

method, however, also decreased the polar lipid content, i.e. 1.6-1.7 times lower than that

without washing. Although salt solution was used for washing in this method, loss of some

polar lipid in the aqueous phase is still a concern.

105

r com

poun

ds (g

/lOO

g lip

id)

Lipi

d ext

ract

ed (g

/1 O

Og f

ish)

1:10 1:20 1:30 1:40 1:50Tissue solvent ratio (w/v)

Figure 4.2 Effect of tissue: solvent ratio on the amount of lipid extracted by theHara and Radin (1978) method

A : no washing ■ : washing

Tissue:solvent ratio (w/v)

Figure 4.3 Effect of tissue:solvent ratio on the level of polar compounds in lipids extracted by the Hara and Radin (1978) method

A : no washing ■ : washing

106

ou,Ph

Tissue:solvent ratio (w/v)

Figure 4.4 Effect of tissue: solvent ratio on protein contents in lipids extracted by the Hara and Radin (1978) method

A : no washing ■ : washing

Different proportions of tissue to solvent resulted in different amounts of lipid

recovered. The lowest proportion of tissue to solvent (1:50 w/v) produced the highest

amount of lipid and, conversely, produced the lowest amount of polar lipid (Fig. 4.5). No

differences (p>0.05) in the amount of protein contaminant extracted with different ratios of

tissue to solvent were revealed (Fig. 4.5). The tissue to solvent ratio of 1:10 (w/v)

produced the highest amount of polar compounds, which was not significantly different

(p>0.05) from that extracted by a ratio of 1:20 (w/v), while other proportions produced

lower amounts of polar compounds. A significant increase (p<0.05) in lipid recovery was

observed for the ratios of tissue to solvent of 1:40 and 1:50 (w/v), however, with these

proportions, difficulty in extraction occurred especially during the washing step. Since the

proportion of tissue to solvent of 1:30 (w/v) was not significantly different (p>0.05) from

that of 1:20 (w/v) in respect of the amount of lipid extracted, while in the respect of polar

107

lipid, the latter proportion produced higher level, the tissue to solvent ratio of 1:20 (w/v)

was chosen for comparative studies.

o 12

o 0

Tissue:solvent ratio (w/v)

Figure 4.5 Lipid yield, polar compounds and protein contents of lipid extracted by the Hara and Radin (1978) method with washing

protein (mg/g lipid): A tot. lipid (g/lOOgfish): ♦ polar lipids (g/lOOg lipid): ■

Based on the higher level of impurities present in, and the cloudy appearance of the

extracted lipid, even when centrifuged, the inclusion of a washing step in the Hara and

Radin (1978) method is recommended. Based on the lipid recovery, and the level of protein

contaminants and polar lipids, a tissue to solvent ratio of 1:20 (w/v) was considered as the

optimum. Although higher proportions of tissue to solvent (1:40 w/v and 1:50 w/v)

extracted more lipid, they were also difficult to handle especially during the washing step

and solvent evaporation, when a considerable amount of lipid should be extracted.

Therefore, such ratios are applicable only for the extraction of very small tissue samples.

Moreover, the high proportion of solvent to tissue is unreasonable in terms of solvent cost

108

and environmental burden.

In this study, the Bligh and Dyer (1959) method was compared to the Hara and

Radin (1978) method, including the washing step with a ratio of tissue to solvent of 1:20

(w/v) in terms of lipid recovery, protein contaminant, polar lipid recovery and stability of

lipid.

The amount of lipid recovered in the Bligh and Dyer (1959) method was lower than

that by the Hara and Radin (1978) method, while the level of protein contaminants and

polar lipids was higher (Table 4.2). This might be due to differences in the tissue to solvent

ratio between the two methods. The proportion of tissue to solvent in the Bligh and Dyer

(1959) method was 1:4 (w/v), while in the Hara and Radin (1978) method it was 1:20

(w/v). The lower solvent to tissue ratio results in more water, but less non-polar solvent in

the solvent system, thus reducing the quantitative nature of the extraction, while increasing

the amount of non-lipids in the extract (Nelson 1991).

Table 4.2 Amount of lipid, protein and polar lipid content* recovered by the Bligh and Dyer (1959) and Hara and Radin (1978) methods

Method Lipid Protein Polar lipid(g/lOOg fish) (mg/g lipid) (g/lOOg lipid)

Bligh and Dyer (1959) 4.32 ±0.04 a 1.04 ± 0.06 c 14.44 ± 0.03 c

Hara and Radin (1978) 4.93 + 0.09 b 0.33 + 0.02 d 14.25 + 0.05/

Values followed by the same letter/s within a column are not significantly different (p>0.05) No correlation among letters in different columns * = Results are means of 3 determinations from 2 replicates

The weight of lipid extracted by both the Hara and Radin (1978) and the Bligh and

Dyer (1959) methods increased during accelerated oxidation at 60°C, due to oxygen uptake

resulting from autoxidation. Lipid extracted by the Hara and Radin (1978) method

adsorbed oxygen faster than that extracted by the Bligh and Dyer (1959) method as

109

indicated by a marked weight gain, whereas lipid extracted by the Bligh and Dyer (1959)

method did not increase markedly until 7 days of incubation (Figure 4.6). The time required

to achieve 1% weight gain during accelerating oxidation at 60°C was used to compare the

relative resistance to autoxidation (Ke et al. 1977, Kaitaranta and Ke 1981). As shown in

Figure 4.6, the time to achieve 1% weight gain of lipid extracted by the Hara and Radin

(1978) method in this experiment was only 2-3 days, while lipid extracted by the Bligh and

Dyer (1959) method required 9-10 days.

Incubation time (days)

Figure 4.6 Weight gain of sardine lipids during autoxidation at 60°C ▼ :Hara and Radin (1978) ■ : Bligh and Dyer (1959)

The fluorescent products formed during lipid oxidation in biological tissues were

shown to be produced as a function of time of peroxidation (Dillard and Tappel 1971).

This was also revealed in the present experiment (Figure 4.7) where the amount of

fluorescent products in sardine lipids undergoing autoxidation at 60°C increased with

increase of time during accelerated oxidation. Lubis (1989) reported similar results for

dried salted sardines.

110

The increase of fluorescence in the present study also correlated well with the

increase of weight gain during accelerated oxidation, especially with lipid extracted by the

Hara and Radin (1978) method with a correlation coefficient of 0.90, while for lipid

extracted by the Bligh and Dyer (1959) method, the correlation coefficient was only 0.56.

Incubation time (days)Figure 4.7 Fluorescene of sardine lipids during autoxidation at 60°C

▼ :Hara and Radin (1978) ■ : Bligh and Dyer (1959)

Consistent with the more rapid increase in weight of lipid during accelerated

oxidation, lipid extracted by the Hara and Radin (1978) method produced a higher amount

of fluorescent products than that extracted by the Bligh and Dyer (1959) method. This

indicated that lipid extracted by the Hara and Radin (1978) method was less stable than that

extracted by the Bligh and Dyer (1959) method. The different solvent system used might

perform differently in extracting endogenous antioxidants or prooxidants thereby affecting

the stability of extracted lipid. Hara and Radin (1978) suggested that, in contrast to

chloroform/methanol, hexane/isopropanol extracted very little pigment from the sample.

The ability of chloroform/methanol to extract more pigments such as carotenoids, one of

111

the biological pigments available in seafood (Haard 1992), suggests that the stability of

lipids extracted by chloroform/methanol solvent might be due to such carotenoids as

antioxidants (Hultin 1992a). Tocopherols (Ke et al. 1977, Kinsella 1988, Chan and Decker

1994) are present in significant amounts in fish flesh including sardine (Pigott and Tucker

1990, Izumi 1993). Either the ability of chloroform/methanol solvent or inability of

hexane/isopropanol solvent to extract tocopherol from tissue matrices might contribute to

the different stability of the lipids extracted by such methods. However, the amount of such

antioxidants was not measured in this study, hence, the definite reason of the phenomenon

is still not clear. The higher proportion of solvent to tissue (20:1 v/w) as well as the higher

boiling point of the solvent used in the Hara and Radin (1978) method, required a longer

time to remove the solvent, and this may have affected the quality of extracted lipid.

In comparison with the Bligh and Dyer (1959) method, the Hara and Radin (1978)

method, including the washing step, was comparable in the recovery of lipid and polar lipid,

but the former method extracted more protein than the latter.

Regarding stability, however, the lipid extracted by the Bligh and Dyer (1959)

method was more stable than that extracted by the Hara and Radin (1978) method.

Therefore the Bligh and Dyer (1959) method was used for subsequent experiments.

4.2 Spoilage Pattern of Australian Sardines

The physical characteristics of raw materials used in these experiments are shown in

Table 4.3.

The sensory evaluation of Australian sardines using the demerit point score system

during deterioration at 28-30° C, RH 70-80%, is shown in Fig. 4.8, while the changes in

attributes are presented in Table 4.4.

112

Table 4.3 Physical dimensions of fresh Australian sardines

Dimension* Experiment 1 Experiment 2 Experiment 3Length (cm) 17.0- 18.5 14.0- 16.0 14.5 - 16.5Width (cm) 2.5 -3.0 2.4-2.7 2.5 -2.8Thickness (cm) 1.5 -2.0 1.3 - 1.8 1.5 - 1.7Weight (g) 40.0- 50.0 25.0 - 34.0 25.0 - 35.0* Range for 10 fish

c 20

B 10

Incubation time (h)Figure 4.8 Changes of demerit point score of sardines incubated at 28-3 0°C

(From 3 replicates, Y = 1.93 + 5.68, r = 0.99)

Fresh sardines used in these experiments had an average demerit point score of 7

which was classified as good quality (Branch and Vail 1985). They had a bright

appearance, firm skin, scale and belly, and had gills with characteristic colour and smell

(Table 4.4). Slight changes in gill colour (darker) and belly (slightly softened and

discoloured) were detected after 4h storage. Similar changes were also observed for

Indonesian oil sardines (Sardinella longiceps) incubated for 4h at 25-26.5°C (Nasran and

Arifudin 1982). After 6h, the belly started to soften and discolour while, although the tissue

113

also softened, the appearance was still bright and other attributes, such as skin, scale and

vent were still in normal condition. This was in agreement with Bremmer et al. (1985) who

noted that, different from the spoilage pattern of fish stored in ice, at which the noticeable

Table 4.4 Changes in fish attributes during deterioration

Incubation time (h) Demerit point score (max. 3 9)

Changes in attributes

0 7 Appearance bright; skin and scale firm; no slimeEyes clear, normal in shape, no bloodGills characteristic in colour, no mucusBelly firm and no discolouration; cavity opalescent, blood red

4 13 Appearance bright; skin and scale firm; slightly slimyEyes clear, normal in shape, no bloodGills slightly dark in colour, fresh, no mucus Belly soft, slight discoloration; cavity opalescent, blood redVent normal, smell neutral

6 16 Appearance bright; skin and scale firm; slightly slimyEyes slightly cloudy, slightly sunken, slightly bloodyGills slightly dark in colour, fishy, mucus moderateBelly soft, slight discoloration; cavity opalescent, blood dark red

8 21 Appearance bright, skin and scale slightly loose, slightly slimyEyes slightly cloudy, slightly sunken, slightly bloodyGills slightly dark in colour, fishy, mucus moderateBelly soft, moderate discoloration; cavity greyish, blood dark redVent slightly open, smell fishy

cont....

114

Table 4.4 (continued)

Incubation time (h) Demerit point score (max.39)

Changes in attributes

10 25 Appearance slightly dull, skin soft and scale slightly loose, slimyEyes slightly cloudy, sunken, very bloodyGills slightly dark in colour, stale, mucus moderateBelly burst, moderate discoloration; cavity greyish, blood dark redVent slightly open, smell fishy

12 29 Appearance slightly dull, skin soft and scale loose, slimyEyes slightly cloudy, sunken, very bloodyGills slightly dark in colour, stale, mucus moderateBelly burst, excessive discoloration; cavity yellow-brown, blood dark redVent slightly open, smell fishy

14 34 Appearance dull, skin soft and scale slightly loose, very slimyEyes cloudy, sunken, very bloodyGills very dark in colour, spoiled, mucus excessiveBelly burst, excessive discoloration; cavity yellow-brown, blood brownVent slightly open, smell spoiled

characteristics of changes occurs in the clarity and shape of eyes, the first sign of

deterioration of fish stored at ambient temperature was the decrease of stiffness, while the

development of mucus and odour in the gills was detected later. Gorczyca et al. (1985)

noted that, after 6h incubation at 37°C, Australian rainbow trout (Salmo gairdneri) lost its

sheen and the texture softened, while no off odour was detected. The average demerit

point score of sardines in this experiment after 6h incubation was 16. Although fish having

this score are categorised as good quality according Branch and Vail (1985), it is close to

the fair quality border.

115

When the incubation continued to 8h, the belly, skin and flesh continued to lose

stiffness and colour. The eyes were slightly cloudy, sunken and bloody, while the scales

were loose. Indonesian oil sardines (S. longiceps) showed similar changes after incubation

for the same period at 25-26.5°C (Nasran and Arifudin 1982). Sardines in the present

experiment had an average demerit point score of 21 which could be categorised as fair

quality. Spoilage was detected after lOh incubation at which the smell of the gills and

viscera was stale, the belly started to burst with moderate discoloration and the belly cavity

turned to greyish. The demerit point score of sardines at this time was 25 and was classified

as poor according to the classification proposed by Branch and Vail (1985). Australian

rainbow trout incubated at 37°C developed off-odours after 9h, which was evidenced as a

sour odour, followed by a putrid (ammoniacal) odour after 1 lh, at which time the fish was

rejected in a sensory evaluation (Gorczyca et al. 1985). In addition, at 11 h the muscle of

Australian rainbow trout was mushy and considerable drip was produced. Indonesian oil

sardines (S. longiceps) were rejected after lOh incubation at 25-26.5°C at which time

rancid and off odours were detected, especially that of mucus which was also sticky (Nasran

and Arifudin 1982).

The quality of sardines in the present study deteriorated after incubation times of 12

and 14h, with average demerit points score of 29 and 34, respectively. Excessive

discoloration and belly burst were noticeable, with yellow-brownish belly cavity. The

appearance became dull, the skin and flesh became very soft and the vent was smelly. In

pelagic species such as sardines, the belly is the most vulnerable part of the body, and it is

the first part of the body to deteriorate (burst), especially in heavy feeding fish. This is

especially due to the activity of the gut enzymes attacking the gut and the surrounding

tissue (Connell 1980, Huss 1995, Bonnell 1994).

The demerit point score was highly correlated with incubation time (r = 0.99), and

116

can be represented by Y = 1.93X + 5.68, where Y is the demerit point score and X is the

incubation time (h). Bremner et al. (1985) assessed the sensory deterioration of four

tropical fish (threadfin breamJNemipterus peronii, longspine seabream/^rgyro/?5 spinifer,

painted sweetlipIPlectorhynchus pictus and snapper-like fishJLutjanus vittus) at ambient

temperatures (25-26°C) for up to llh. They found that the slope of linear correlation

between demerit point score and storage time was in the range of 1.14-1.51. This indicates

that Australian sardines in the present experiment spoiled at a higher rate (1.93 demerit

points/h) than the fish observed by Bremner et al. (1985).

Incubation time (h)

Figure 4.9 Changes of total volatile base nitrogen of sardines incubated at 28-30°C (From 3 replicates, Y = 10.98e° °7X, r = 0.97)

The total volatile base nitrogen (TVB-N) level increased with the increase of

incubation time (Fig. 4.9). Before incubation, Australian fresh sardines had an average

TVB-N of 13 mg%N, which increased to 16 mg%N after 6h of incubation. When the sign

of spoilage was detected (after lOh incubation), the TVB-N level was 20 mg%N. Further

117

incubation of sardines to 12h and 14h in the present study increased the TVB-N levels to 23

mg%N and 31 mg%N, respectively.

The correlation between TVB-N level and incubation time was exponential and can

be represented by Y = 10.98e0 07X (r = 0.97), where Y is TVB-N (mg%N) and X is

incubation time (h). Exponential correlation between TVB-N and incubation time was also

observed by Estrada et al. (1985) for whiting (Sillago maculata) and tilapia (Oreochromis

niloticus) stored at 29°C, 71% RH.

Generally, the limit value of TVB-N for fish acceptability is around 30 mg/100 g

flesh (Sikorski et al. 1990a), however, this value also depends on the species of fish. The

limit of acceptability of cold-water fish in ice was at a level of 30-40 mg TVB-N/lOOg flesh

(Huss 1988). For fatty fish species such as herring and mackerel, the maximum value of

TVB-N for acceptable fish is about 20 mg per lOOg flesh (Sikorski et al. 1990a).

Nasran and Arifudin (1982) found that Indonesian oil sardine {Sardinella longiceps)

kept at ambient temperature (26-28°C) was rejected by panelists (using a hedonic scale)

after lOh, after which time the TVB-N value reached 20 mg%N. In the present experiment,

Australian sardines {Sardinella neopilchardus) showed TVB-N values of 20 mg%N after

lOh incubation at 28-3 0°C and RH 78-80%, supporting the finding of Nasran and Arifudin

(1982). Demersal fish, such as rabbit fish (Siganus sp), was rejected after 18h storage at

Indonesian ambient temperature (28.8-32°C), at which time the TVB-N level was 75 mg%N

(Putro, Saleh and Utomo 1985). Upon incubation at 29°C, 71% RH, whiting {Sillago

maculata) and tilapia {Oreochromis niloticus) had TVB-N levels of 52 mg%N after 12h,

and 15 mg%N after 16.5h, respectively (Estrada et al. 1985). At these levels, both fish

were rejected by a taste panel.

Differences in the detection of spoilage or the rejection of those fish and sardines in

the present study were due to species differences. The rejection time of tilapia, which was

118

longer than the other fish in both studies, was especially due to the lower level of TVB-N

causing less offensive odour. Moreover, similar to rabbit fish, tilapia had harder and more

compact skin than did whiting and sardines. The demerit point score of Australian sardines

after lOh incubation was 25 and it can be graded as poor quality according to Branch and

Vail (1985).

A linear correlation was found between TVB-N levels and demerit point (Figure

4.10), and can be represented with the equation Y = 1.48X - 8.13 (r = 0.95), where Y is

TVB-N (mg%N) and X is the demerit point score. This means that demerit point score,

besides being a measure of quality deterioration of fish, can also be used to predict the

TVB-N level. Demerit point score is also more reliable in predicting the remaining shelf-life

of fish stored at various temperatures (Branch and Vail 1985).

Demerit point scoreFigure 4.10 Correlation between demerit point and total volatile base nitrogen of

sardine incubated at 28-30°C (Y= 1.48X- 8.13, r = 0.95)

The pH of fish during incubation was relatively constant (Fig. 4.11). The average

initial value was 6.1 and slightly increased to 6.2 after 6h. However the pH decreased after

119

lOh incubation, at which spoilage was detected, and increased for the following time of

incubation but the value never exceeded 7.0. The use of pH as an index of spoilage is of

questionable value (Krishnakumar, Hiremath and Menon 1986, Gelman, Pasteur and Rave

1990).

Incubation time (h)Figure 4.11 Changes of pH of sardine incubated at 28-30°C

(Each point is an average of 3 replicates)

In producing dried salted sardines, low quality raw material is commonly used.

However, this does not mean that completely spoiled sardine is suitable to be processed for

dried salted fish, since the belly of spoiled sardines is often burst and the texture of the flesh

is soft and flabby. Nasran and Arifudin (1982) suggested that raw sardine suitable for

drying and salting is that left at ambient temperature (26-28eC) for not more than 7h, while

TVB-N value of raw sardine commonly used by traditional processors for dried salted

product is commonly 20 mg%N (Saleh and Murtini 1982).

Similarly, the lowest quality Australian sardine suitable for dried salted fish

production is preferably not completely spoiled (sardine incubated for 12h or more), but one

120

that shows slight to moderate signs of spoilage. This can be obtained by incubating

sardine at the above conditions for lOh, which will result in a TVB-N value of

approximately 20mg%N in the flesh, a demerit point score of 25 and a poor quality sardine

product. This was in accordance with the raw sardines commonly used by traditional

processors in Indonesia (Saleh and Murtini 1982). In order to examine dried salted sardine

produced from partly spoiled raw material (between fresh and the most deteriorated),

sardines were incubated for 6h to produce a demerit point score of 16 and TVB-N level of

16 mg% N.

4.3 Lipid Characterisation of Raw Materials and Their Dried Salted Products

4.3.1 Chemical analysis

The characteristics of raw material of different qualities are shown in Table 4.5.

The amount of lipid extracted from different quality raw material was not significantly

different. Similarly, there was no significant difference in the amount of lipid extracted from

their corresponding dried salted products.

Table 4.5 The characteristics of fish raw materials

Raw material Demerit point score* TVB-N(mg%N)**

TMA-N(mg%N)**

Fresh 6 ± 0.3 a 111 ±0.00d 0.31 ±0.03 hModerately spoiled 16 + 0.6 b 16.0 ±0.01 e 0.47 ±0.03 iSpoiled 25 ±0.1 c 20.4 ±0.62/ 0.73 ±0.00 /Values followed by the same letters within a column are not significantly different (p>0.05) * Results are means of 5 fish from 2 replicates ** Results are means of 3 determinations from 2 replicates

There was a marked decrease in lipid extracted after processing (Fig. 4.12). This

was expected since oxidation takes place during salting and drying producing compounds

121

which are relatively insoluble in organic solvents due to reaction of fatty acids with other

compounds such as amino acids or carbohydrates. Besides, during salting, a part of the lipid

substances might dissolve in salt solutions that are eliminated together with the moisture.

Syamsiar, Fawzya and Poernomo (1986) noted that, during 12 h of wet salting (brining), the

loss of fat in oil sardines was 23%.

According to Ackman (1992), drying of food matrices might encase lipids in dry

protein or some similar matrix that solvent could penetrate only with difficulty. Moreover,

drying also results in the oxidation of PUFA creating polar compounds of relatively low

solubility in many organic solvents, and reaction of oxidised PUFA with other fatty acids

produces relatively insoluble polymers, while reaction of oxidised PUFA with amino acids

or carbohydrates gives more permanent and insoluble binding of lipid (Ackman 1992).

-a-C

c/5incooo

Fresh Mod. spoiled Spoiled

Raw material quality

Figure 4 .12 Amount of lipid extracted from sardines before and after processing

Hraw material ■ dried salted fish

The amount of lipid extracted from sardines before and after processing differed by

45-55%. This was supported by Nasran et al. (1992) who processed ray into dried salted

122

fish which resulted in a decrease of lipid extracted in the range of 50-84%. The greater

changes in lipid extracted from fish processed by these authors was due to the thinner

slices, resulting in more lipid dripping during salting and more surface contact with oxygen.

Astawan et al. (1994a), who processed skipjack tuna (Katsuwonus pelamis) into dried

salted fish, found that the changes of lipid extracted after processing were in the range of

44-47%, while decrease in lipid extracted after salting and drying of marine catfish {Arms

thalassinus Ruppell) was 23% (Smith et al. 1990). It seems that the decrease in lipid

extracted after salting and drying is dependent on the size of fish such as the thickness of

fillet, and the salting process (time and concentration of salt).

0.25

Fresh Mod. spoiled Spoiled

Raw material quality

Figure 4.13 Level of fluorescence of lipid extracted from sardines before and after processing

HI raw material ■ dried salted fish

Before processing, there were no differences in the level of fluorescence from raw

materials. After processing, the level of fluorescence increased markedly (Fig. 4.13) from

0.02 pg/g fish before processing to 0.21 - 0.24 pg/g fish after processing; a tenfold increase.

123

Similarly, Smith et al. (1990) found that the increase of fluorescence level of marine catfish

after salting and drying was approximately twenty fold, whereas Lubis and Buckle (1990)

showed that the level of fluorescence after salting and drying of Australian sardine

(Sardinops neopilchardus) was 0.3 pg/g fish; a level close to that in the present study. This

clearly indicates that during processing of dried salted fish, oxidation took place and

fluorescent products were produced as a result of the reaction of malonaldehyde with

proteins and other cellular constituents (Logani and Davies 1979, Melton 1983).

Different raw material quality produced different levels of polar compounds in the

extracted lipid (Fig. 4.14). The highest level of polar compounds was present in lipid

extracted from spoiled raw material, while lipids from fresh raw material contained the

lowest level of polar compounds. According to Kaitaranta and Ke (1981), polar lipid

content can be used as an indicator of the oxidation level of fatty fish lipid. The increase in

polar lipid content of 33%, which represented 1% weight gain during an accelerated

oxidation test at 60°C, was used by Ke et al. (1977) as an oxidative index for comparing the

antioxidative effectiveness of various antioxidants. In the present study, the increase of

polar compounds from fresh to moderately spoiled raw materials was only 16%, while that

from moderately spoiled to spoiled and from fresh to spoiled raw materials were 34% and

55.8%, respectively. It seems that the oxidation of lipid occurred during accelerated

spoilage and it is likely that the increase in polar compound was resulted, especially from

the PUFA. According to Frankel (1983) autoxidation of PUFA (linoleate and linolenate)

forms significant amounts of polar and polymeric materials. Similarly, a development of

polar compounds resulting from oxidised marine lipid that occurred during refrigerated

storage at 2°C (Joseph and Seaborn 1982) and after frozen storage (-12°C for 3 months) of

Atlantic salmon (Polvi et al. 1991) was also noted.

124

After salting and drying, the polar compounds in fresh and moderately spoiled fish

increased significantly, while in spoiled raw material the increase in polar compounds was

less marked. This shows that salting and drying also caused extensive autoxidation,

creating polar compounds, supporting the findings of other workers (Ackman 1992, Yankah

et al. 1993). Kaitaranta and Ke (1981) observed an increase in polar compounds of

Atlantic herring (Clupea harengus) and mackerel {Scomber scombrus) during accelerated

oxidation at 60°C.

Fresh Mod. Spoiled Spoiled

Raw material quality

Figure 4.14 Level of polar compounds of lipid extracted from sardines before and after processing

lUraw material ■dried salted fish

In the case of fatty acid composition, no changes in SAFA and MUFA were

observed during accelerated spoilage, but there were significant changes in PUFA

(p<0.05). PUFA of lipid extracted from both moderately spoiled and spoiled raw materials

were lower than that of fresh fish (Fig. 4.15). This suggests that, during accelerated

spoilage of sardines, PUFA suffered oxidation.

The losses in PUFA during the spoilage process seem to be predominated by DHA

(Table 4.6). This is in agreement with Hardy et al. (1979), Joseph and Seaborn (1982) and

125

Polvi et al. (1991) who noted losses of DHA during cold and frozen storage of marine

fish, while a decrease in DHA and EPA of sardine oil incubated at 22°C for one month was

observed by Suzuki et al. (1985). The ratio of PUFA either C22:6 or the sum of C22:6 and

C20:5 to Ci6:o, was used as an index of oxidative rancidity, i.e. PI (Lubis and Buckle 1990,

King, Boyd and Sheldon 1992, Davis et al. 1993), therefore losses of such PUFA

correspond to the decrease in PI (Table 4.6).

Raw material quality

Figure 4.15 Fatty acids of lipid extracted from sardines before and after processing

HI raw material ■ dried salted fish

Substantial losses of PUFA were noted after processing, while the changes of

SAFA and MUFA were negligible (Fig. 4.15). The losses of PUFA after processing were

in the range of 22-29%, with 65-69% contributed by DHA and 15-19% by EPA. It is

interesting to note that the differences in raw material quality did not give any significant

differences in PUFA from the corresponding products. This suggests that oxidation

occurred so extensively, and complex reactions involving lipid and other biologically active

compounds took place due to the presence of salt during salting in combination with heat

126

during drying, that no obvious differences in the level of PUFA of the products were

observed. Although the exact mechanism of salt action is still not clear, the acceleration

effect of salting toward oxidation is evident (Shewan 1955, Castell et al. 1965, Nambudiry

1980, Hultin 1992a), while drying by exposing fish to heat and/or light clearly results in lipid

oxidation (Howgate and Ahmed 1972, Bligh et al. 1988). The present results confirm the

finding of previous workers (Rao and Bandyopadhyay 1983, Maruf et al. 1990, Smith et al.

1990). Smith et al. (1990) noticed 30% loss of PUFA during salting and subsequent drying

of Indonesian salted-dried marine catfish (Arius thalassinus) and other workers found that a

significant loss of PUFA was noted immediately after processing of Indian mackerel

(Rastre/liger kanagurta) (Rao and Bandyopadhyay 1983) and Indonesian dried-salted

mackerel (Rastrelliger kanagurta) (Maruf et al. 1990).

The polyene index, that directly correlated with PUFA and SAFA in this study,

decreased significantly (p<0.05) after processing (Fig. 4.16). The initial PI (before

processing) was in the range of 2.1-2.3, while after processing the value decreased to 1.7.

The later value was close to PI (1.5) found by Lubis and Buckle (1990) who salted and

dried the same fish. Similarly Yankah et al. (1993), who worked with dried salted Japanese

mackerel (Trachurus japonicus), concluded that a decrease in the proportion of total DHA

and EPA to C 16:0 (from 0.93 to 0.78) occurred during processing. The decrease of PI

after processing in this study was coincident with the decrease in PUFA after processing

(Table 4.6). This is expected, since, during salting and drying, considerable losses of PUFA

occurred, while the level of SAFA including C 16:0 remained constant. Thus the

proportion of PUFA and C 16:0 known as PI decreased. As in the case of PUFA, PI of

products prepared from different quality of raw materials was not significantly different

(p>0.05).

127

Fresh Mod. spoiled Spoiled

Raw material quality

Figure 4.16 Polyene index of lipid extracted from sardines before and after processing

i!raw material Bdried salted fish

4.3.2 Headspace analysis

The volatile headspace profiles from lipid extracted from different qualities of raw

material are illustrated in the chromatogram in Fig. 4.17 and Fig. 4.18. The differences in

the profile of chromatograms among different quality raw materials are demonstrated, at

which the intensity of the peaks increased with the decrease in raw material quality.

Quantitative analysis was not based on absolute concentration, but rather roughlv based on

relative proportions of the volatile compounds to internal standard (2-heptanone) that was

added to each injected sample Naturally, some errors might occur due to different partition

coefficients of either the internal standard or individual compounds, since the response

factors both of the internal standard and the reference compounds were not determined in

this study due to unavailability of some of the reference compounds.

128

Table 4.6 Fatty acid compositions of different quality raw materials and the corresponding dried salted products

Fatty Raw material Dried salted fish fromacid Fresh Moderate­

ly spoiledSpoiled Fresh Moderate­

ly spoiledSpoiled

C 14:0 2.5 2.3 2.2 2.1 2.1 2.3

C 16:0 14.3 14.2 14.4 13.3 13.3 13.2

C 18:0 3.0 2.9 3.4 3.4 3.4 3.4

C 20:0 0.1 0.1 0.1 0.1 0.1 0.1

SAFA 19.6 19.6 19.8 18.9 18.9 19.0

C 16.1 n-9 2.3 2.3 2.1 2.1 2.1 2.1

C 18:1 n-9 3.1 3.0 3.0 3.1 3.1 3.1

C 18:1 n-7 1.3 1.2 1.2 1.4 1.4 1.4

C 20:1 n-9 0.1 0.1 0.1 0.1 0.1 0.1

C 20:1 n-7 0.0 0.0 0.1 0.1 0.1 0.1

MUFA 6.8 6.6 6.5 6.7 6.7 6.7

C 18:2 n-6 0.7 0.7 0.7 0.8 0.8 0.8

C 18:3 n-3 0.4 0.4 0.4 0.3 0.3 0.3

C 18:4 n-3 0.6 0.6 0.5 0.4 0.4 0.5

C 20:2 n-6 0.1 0.1 0.1 0.1 0.1 0.1

C 20:3 n-6 0.1 0.1 0.1 0.1 0.1 0.1

C 20:4 n-6 1.4 1.4 1.4 1.1 1.1 1.1

C 20:4 n-3 0.2 0.2 0.2 0.2 0.2 0.2

C 20:5 n-3 6.5 6.1 5.9 4.4 4.4 4.3

C 22:5 n-6 1.0 1.0 1.0 0.8 0.8 0.7

C 22:5 n-3 0.7 0.6 0.6 0.5 0.5 0.5

C 22:6 n-3 24.4 22.3 22.7 17.1 17.1 16.5

PUFA 36.1 32.9 32.8 25.7 25.7 25.3

PI 2.3 2.1 2.1 1.7 1.7 1.7

129

Figure 4.17 C

hromatogram

of sardine lipid extracted from

different quality of raw m

aterials

130

Figure 4.18 C

hromatogram

of sardine lipid extracted from dried salted sardine

prepared from different quality of raw

materials

131

In order to determine the odour characteristics of the individual aroma components,

sniffing of GC effluents of the fish lipid was performed. In this case the nose acts as a

detector, hence the objective assessment carried out by GC FID can be correlated with

subjective sensory information. Aromagrams were obtained from the lipid extracted from

dried salted fish prepared from spoiled raw material since this contained all volatile

compounds that were representative of all the other lipids.

In this study, not all peaks could be characterised, due to the fact that some

components were too small and not separated well. Moreover some large peaks might not

give any odour, while a small peak could give a strong aroma. This is due to the different

sensitivity of the two detectors, i.e. FID and nose to volatile compounds that were

separated by GC. Moreover, sniffing fatigue created difficulty in detecting the aroma of

compounds. Hence, in order to get representative aromagram results, three sniffing runs

were carried out. Further identification of each volatile compound from the aromagram by

GC-MS was not carried out in this study, however, characterisation and identification as

well as estimation of volatile compounds in fish and/or fish lipid published by previous

workers is given.

Based on the odour quality resulting from the sniffing technique, different characters

starting from pleasant aromas, such as fruity and floral, to unpleasant aromas, such as burnt

and fishy, were detected. The aromagram representing the odour character from individual

peaks is shown in Figure 4.19. These odour characteristics were classified into five major

groups according to Lin’s classification (1994), namely, green, sweet, oxidised, pesticide­

like and others. Each group consists of minor notes of aroma characteristics as shown in

Table 4.7.

132

Table 4.7 The major ami minor classification of aroma of lipids from dried salted sardine

Green Sweet Oxidised/ Pesticide- Othersrancid like

apple sweet oily benzene nuttycucumber floral fishy meatymelon deep fried woodygrass painty cardboardstink bug burnt musty

earthydirty sockcamphormedicinalastringentmushroom

The total amount of volatile compounds of lipid detected by the sniffing technique increased

(both in lipid extracted from different quality raw materials and their corresponding dried

salted fish) with decreasing raw material quality (Fig. 4.20). This was in coincidence with

the profile of their aromagrams (Fig. 4.17 & Fig. 4.18). The volatile compounds of lipid

extracted from raw materials increased gradually with the decrease in raw material quality,

however, they increased sharply in lipid extracted from dried salted fish prepared from

spoiled raw material. This suggests that the spoilage process combined with salting and

drying enhanced the oxidation process, resulting in an increased amount of volatile organic

compounds. Lipid extracted from dried salted fish suffered more oxidation during salting

and drying, and the formation of volatile compounds in the latter was also due to the

reaction of oxidation products with other food components such as protein, amino acids

and carbohydrates known as Maillard reactions (Whitfield 1992).

133

OnQ

*un« »

fr n'

ftnoxum

#«••*.

Aj*

*u

^ tx^ooo M

C’U

Ai»*ui/r»»u< p«

«m »«*

i^

1501 AvJfp »t» *••)M

O»>

UI

*/••*

•

vmipj^q •#•*(

MM

«~je to«o«. •*-n

pe>l-J a®

** •**,. • *,

A*S«V A#o •»•*•»

o•5econTJo3SOUu e 4. |9 Aromagram of lipid extracted from dried sal

prepared from spoiled raw material

Con

cent

ratio

n (1

0'6)

1600

1400

1200

1000

800

600

400

200

0

Green

i—

Sweet

Oxidised Pesticide-like

Others

FRESH MOD. SPOILED SPOILEDFRESH MOD SPOILED SPOILED

Figure 4.20 Aroma characters of sardine lipids extracted from different quality fish and thecorresponding dried salted products

Raw material Dried salted product

135

The levels of the major character notes slightly increased with the decrease in quality

of raw material, or tended to be constant except for that of lipid extracted from dried salted

fish prepared from spoiled raw material. There was no significant difference (p>0.5) in the

level of major aroma types characterised from lipid extracted from both fresh and

moderately spoiled fish.

Green and oxidised character notes seem to be present in higher levels in lipid

extracted both from different quality of raw materials as well as their corresponding dried

salted fish (Fig. 4.20). In general, the level of major character notes from extracted lipid in

this experiment increased after processing of raw materials into dried salted product, except

that the green character note decreased even in lipid extracted from spoiled raw material.

Although a green aroma is also produced during autoxidation, this flavour is also

accompanied by compounds contributing to general oxidised flavours which are primarily

responsible for burnt oxidised-fishy flavours. Therefore, in more oxidised lipids, the green

aroma reflecting fresh-fish quality diminishes, while burnt, cod liver flavours are enhanced

(Karahadian and Lindsay 1989a). This phenomenon may explain the results of this present

study.

It has been documented that green, melony, grass-like flavours have been

encountered in fish (Hsieh et a/. 1989, Karahadian and Lindsay 1989a, Stansby 1971).

Oxidation of PUFA has been believed to be initiated by enzymatic reactions

(lipoxygenases) resulting in hexanal, unsaturated C8 or C9 alcohols, and unsaturated C9

aldehydes as the end products that are responsible for green, melon-like and green

cucumber notes in fish oils (Karahadian and Lindsay 1989a). According to Karahadian and

Lindsay (1989b), trans,cis-2,6-nonadienal was primarily responsible for the green, fresh

fish-like flavour observed soon after deodorisation of fish oil while compounds responsible

for the heavier green odours, such as green stinkbug and plant-like odours, were trans-2-

136

hexenal, l,5-octadien-3-one and low concentrations of trans,trans,c/5-2,4,7-decatrienal.

Stansby (1971) noted that a green grass-like flavour in menhaden oil was similar to that of

c/s-3-hexen-l-ol. Although green flavour resulted predominantly from an initial oxidation

of PUFA by lipoxygenases, the green flavour was still detected in the later stage of flavour

development in autoxidising fish oil that was commonly accompanied by fishy aroma

(Karahadian and Lindsay 1989b), therefore, it was not surprising that green aroma was

present in all lipids extracted from different quality of raw materials as well as their

corresponding dried salted products. Moreover l-penten-3-ol, contributing to grassy,

green flavours was reported to be the main volatile arising ffom enzymic and/or non-

enzymic oxidation in sardine (Kawai 1996).

The characteristic sweet aroma also occurred at a significant level in lipid extracted

from different quality of raw materials, although the levels were lower than that of the green

aroma. Josephson et a/. (1987) reported that the sweet aroma in refrigerated fish seemed to

be contributed by short-chain alcohols. They noted further that medium chain alcohols,

such as 3,6-nonadien-l-ol, might contribute to the heavy sweet note such as a sweet melon­

like character. Aldehydes such as decanal were also reported to be related to sweet, green

fruity notes in crude menhaden oil, while almond, a sweet fruity note detected in the same

oil, was contributed by benzaldehyde (Hsieh et a/. 1989). These aldehydes were also

obtained in cod liver oil (Karahadian and Lindsay 1989b), whereas in heated sardine,

benzaldehyde was present in significant level (Komi, Kawai and Ishida 1992). The level of

sweet character in volatile compounds increased significantly in lipid extracted from both

moderate and spoiled raw materials.

As was the case for green character notes, the major notes of the oxidised aroma

included fishy, cod liver oil-like, burnt, fried dried salted fish, and a painty aroma was also

present in high level. According to Meijboom and Stroink (1972), the fishy off-flavour

137

occurring in strongly autoxidised oils containing linolenic acid or n-3 fatty acids was due to

trans,trans, c/5-2,4,7-decatrienal They noted that this aldehyde was responsible for the cod

liver-like flavour. However, Ke et al. (1975) suggested that fishy flavour in oxidised lipid

of mackerel was probably not only due to decatrienals, but also due to the presence of a

complex mixture of carbonyl compounds, while Swoboda and Peers (1977) reported that

compounds having the 2,4-dienal functional group contributed to the fishy taint. In

oxidising cod liver oil, a high concentration of decatrienals (400 ppb) imparted distinct

bumt/fishy flavours (Karahadian and Lindsay 1989b). This aldehyde was also obtained in

autoxidised crude menhaden oil (Hsieh et al. 1989).

In the present study, fishy aroma was clearly detected from the GC effluent through

the sniffing port. However, it cannot be stated with certainty what compounds were

responsible for such aroma, since confirmation and identification using GC-MS were not

carried out. Hsieh et al. (1989) detected decatrienal in crude menhaden oil and this

aldehyde emerged at the end of the chromatogram. In this study, fishy aroma was also

detected in small peaks at the end of the chromatogram, i.e. peaks eluted at 108, 109, and

115 min (Fig. 4.19). Considering that the method used to analyse volatile compounds in this

study was adopted from the method of Hsieh et al. (1989), it could be speculated that such

trienals might be present in autoxidised sardine lipid in this study.

The painty taint in oxidised fats was reported to be caused by aldehydes such as

pent-2-enal (Saxby 1982), trans-but-2-enal (Hsieh et al. 1989), hexanal, 2,4-heptadienals

and 2,4-decadienals (Karahadian and Lindsay 1990). According to McGill et al. (1974),

painty off-flavour in cold stored cod was also contributed by hept-4-enal. It was also

reported that rancid catfish oil contained hexanal, 2,4-heptadienals and 2,4-decadienals

(Angelo, Dupuy and Flick 1988) and heated sardines contained hexanal and c/s-4-heptenal

in the volatile compounds (Komi et al. 1992). For the above reasons, the painty aroma

138

detected in oxidised sardine lipids in this experiment might be due to the presence of these

kind of aldehydes.

Regarding the deep fried aroma detected in this present study, it seems that this

correlated with the decadienals present in oxidised lipids as reported by Saxby (1982) and

Love (1996) while Triqui and Reineccius (1995a,b) noted that in addition to decadienals,

deep fried fat flavour in cured anchovy was also due to nonadienals and octadienals.

The unpleasant pesticide-like or petroleum-like aroma detected in sardine lipid in

this experiment may be contributed by aromatic hydrocarbons. The origin of these

compounds was uncertain, but could be from the aquatic environment. Aromatic

hydrocarbons, especially the alkylbenzenes containing nine to ten carbons from

contamination by petrochemicals, was suggested to be responsible for such odours

(Reineccius 1991). These off-flavour compounds were also obtained in crustacean seafood

caught in Australian coastal waters (Whitfield and Freeman 1983), while a flavour

resembling kerosene that occurred in Australian mullet was suggested to arise from water

pollution ( Valq et al. 1970).

Another contributor to petroleum-like off flavour was proposed by Sato et al.

(1988), who reported that one strain of yeast, i.e. Torulopsis Candida (Saito) Lodder 1922

produced styrene that was responsible for the petroleum-like off flavour.

The presence of hydrocarbons and benzene derivatives in some foods could be due

to thermal decomposition of carotenoids, producing methyl naphthalenes, toluene and

xylene (Pippen, Mecchi and Nonaka 1969). Besides found in fish oils (Lin 1994), the

aromatic hydrocarbons were also identified in crayfish waste (Tanchotikul and Hsieh 1989),

boiled crayfish tail meat (Vejaphan, Hsieh and Williams 1988), crabmeat (Matiella and

Hsieh 1990) and cooked chicken meat (Nonaka, Black and Pippen 1967). A petroleum­

like aroma detected in sardine lipid in this study was possibly due to environmental

139

pollution, as in the case of mullet and crustaceans caught on the Australian coast and/or

from decomposition of carotenoid compounds, since these biological pigments, commonly

available in seafood, might also be found in sardines.

Other volatile aromas prevalent in sardine lipids include meaty, woody, camphor and

dirty sock, and are listed in Table 4.7. Meaty aroma was characterised in the area of

retention time 71 to 73 min. Sulphur-containing compounds such as thiazoles, thiophenes

and methylthiopropanal were reported to exhibit a meaty aroma in marine Crustacea

(Vejaphan et al. 1988, Tanchotikul and Hsieh 1989, Cha, Cadwallader and Baek 1993) as

well as methyl mercaptan and acrolein in the fat of cooked poultry (Pippen et al. 1969).

These sulphur-containing heterocyclic compounds were derived from the interaction

between unsaturated fatty acids or their oxidation products with sulphur-containing amino

acids during heating, while saturated and unsaturated aldehydes with 6-10 carbon atoms are

likely contributors to meaty aroma in all cooked meat (Love 1996). A separate study by

Vercellotti et al. (1992) showed that aroma of boiled beef perceived in roasted peanuts has

been identified as hexadecane.

The characteristic woody aroma was also detected in some areas of the GC

aromagram. Alcohols such as 2-decanol (Tanchotikul and Hsieh 1989) and 2-

butoxyethanol (Vejaphan et al. 1988) are the compounds that may be responsible for the

woody aroma.

The musty aroma that appeared in the sniffing runs in this study was also detected in

crude menhaden oil (Hsieh et al. 1989), whereby a musty aroma with a citrus topnote and

musty, waxy floral note were contributed by nonan-2-one and ?ra/7s-oct-2-enal,

respectively. The substituted pyrazines such as 2-methoxy-3-isopropylpyrazine and 2,3,5-

trimethylpyrazine were also found to be primarily responsible for the musty odour (Saxby

1982, Tanchotikul and Hsieh 1989). Naturally occurring musty, earthy and muddy flavours

140

obtained in pond-culture fish were reported to be caused by algae, mainly blue-green algae

producing geosmin (trans-1, 10-dimethyl-/ra«s-9-decanol), which has a distinct earthy-

musty note, and 2-methylisobomeol (MIB) which has a muddy or lagoon blue-green algae

odour (Lovell 1983, Johnsen et al. 1992).

Sweaty socks and dirty socks are common off-flavours derived from lipid and/ or

lipid-containing foods. This aroma perceived in the sniffing run of sardine lipids in this

study might be caused by some fatty acids (as obtained in crude menhaden oil). According

to Lin (1994), butanoic acid, isobutanoic acid and hexanoic acid were the major

compounds imparting dirty and sweaty sock aromas. Love (1996) reported that a dilute

solution of some fatty acids, especially shorter, branched-chain acids, had a cheese-like

aroma. Acids were formed by direct thermal decomposition of triacylglycerols, hydrolysis

of the ester linkages and autoxidation of aldehydes (Grosch 1982).

Other off-flavours detected in sardine lipid were camphor-like or naphthalene-like

and undesirable medicinal aromas. The latter aromas were characterised by other workers

in fish oils and crustacean by-products, and they noted that phenol and xylene contributed to

the medicinal aroma (Hsieh et al. 1989, Vejaphan et al. 1988, Tanchotikul and Hsieh 1989,

Cha et al. 1993), while indole has been reported to contribute camphor-like or

naphthalene-like aroma (Chung and Cadwallader 1993). Kawai (1996) reported that indole

and methylindole were found in heated sardine and together with l-penten-3-ol and 4-cis-

heptenal contributed to the odour of sardine, while in heated cod, indole could more affect

the odour quality than the aldehyde. Phenol and 4-methylphenol were also obtained in

heated sardine (Kawai 1996).

141

4.4 Characterisation of dried salted fish flavour

4.4.1 Chemical analysis

Dried salted fish processed from different quality of raw material have proximate

composition as follows:

Table 4.8. The characteristics of dried salted fish

Dried salted fish prepared from

Characteristic Fresh fish Moderately spoiled Spoiled

Moisture (% wb) 45.51 ± 0.06 nr 43.98 ± 0.20 b 46.31 ±0.18 c

Protein (% wb) 35.59 ± 0.36 a 37.79 ±0.43 b 33.90 ±0.00 c

Lipid (% wb) 0.98 ± 0.00 a 1.13 ±0.06 nr 1.09 ± 0.18 nr

Ash (% wb) 18.34 ± 0.14 nr 18.88 ±0.01 nr 18.45 ±0.03 nr

Salt (% wb) 14.82 ±0.22 a 14.82 ± 0.37 nr 14.65 ±0.35 a

Aw 0.77 ±0.00 nr 0.77± 0.00 nr 0.77± 0.00 a

Values followed by the same letters within a row are not significantly different (p>0.05) Values are mean + SD from 2 replicates No correlation among letters in the same column

As presented in Table 4.8, the moisture content of dried salted fish are in the range

of 44 - 46% wb and the aw were all 0.77, which were close to those of Lubis (1990) who

studied the same products. Water activity of Indonesian dried salted fish from 24 species

were in the range of 0.65-0.79 (Pitt 1995). Although this level could prevent the growth of

pathogenic and putrefactive bacteria, halophilic bacteria and moulds could still grow (FAO

1981). In relation to lipid oxidation and enzyme activity, higher reaction rates occur in

the range of 0.3-0.8 aw (Olley, Doe and Heruwati 1988). The aw values of dried salted

142

fish in this experiment were 0.77, this clearly indicates that oxidation might occur in dried

salted fish in this study. There was no significant difference (p>0.05) in aw among dried

salted fish prepared from different raw material quality, although the moisture content was

different. It seems that different quality raw materials did not give any effect on salt uptake

as shown in Table 4.8 and consequently there was no difference in aw of products.

Besides being used to specify the freshness of fish, TVB-N is also used to asses

quality of fish products including dried salted fish and the levels of 100-200 mg TVB-

N/lOOg fish have been recommended for a variety of salted and dried fish (Connell 1980).

The TVB-N levels in this study that did not exceed 41 mg/lOOg fish (Fig. 4.21) and were

lower compared to that of recommended levels. However, the levels of TVB-N are likely

to depend upon the species and processing methods.

50

Fresh Mod. spoiled Spoiled

Raw material quality

Fig. 4.21 Total volatile base nitrogen of dried salted fish

Ulraw material ■ dried salted product

143

A number of dried salted fish including dried morwong (Nemadactylus

macropterus) (Wootton and Ismail, 1986), fringescale sardines (Sardinella fimbriata)

(Poernomo et al. 1988) and skipjack tuna (Katsuwonus pelamis) (Astawan et al. 1994b)

have been reported to contain TVB-N that were close to the levels found in this study.

Higher levels of TVB-N (76-189 mg/lOOg fish) were found in laboratory dried salted rays,

but these levels were still lower than those of commercial products, i.e. 559 mg/lOOg fish

(Ariyani et al. 1993). This might be due to the high concentration of urea which was

degraded into ammonia (Nasran et al. 1992). These levels of TVB-N in dried salted fish

prepared from fresh fish in this study did not differ significantly from the same products

prepared from moderately spoiled material, while those from spoiled raw material were

significantly higher (p<0.05).

As in the case of TVB-N, TMA-N has also been used to assess the quality of dried

fish. The levels of TMA-N in the present study were much lower when compared to the

level of TVB-N (Fig. 4. 21 and 4. 22). This is expected since TVB-N is composed of

ammonia, TMA-N, as well as DMA and methylamine (Sikorski et al. 1990a). Figure

4.22 shows the levels of TMA-N in raw materials that increased significantly with the

decrease in fish freshness. This suggests that during spoilage, TMA-0 that is commonly

present in marine fish such as sardines was degraded by either bacterial enzyme or

endogenous enzyme actions into TMA-N (Ashie et al. 1996).

After processing, the levels of TMA-N increased markedly (Fig. 4. 22). This might

be due to the degradation of TMA-0 during drying as reported by Colby et al. (1993). The

authors noted that due to the instability of TMA-0 to heat, concentrations of TMA-N

increased greatly after drying. The authors stated further that thermal degradation of TMA-

O is promoted by non-enzymatic processes, as also reported by Hultin (1992b). Although

144

there is a standard limit for TMA-N level in fresh fish, little information is available for the

standard limit of TMA-N for fish products including dried salted fish. Each product has a

different level of TMA-N when rejected. As an illustration, 10-15 mg TMA-N /100g in

aerobically stored cod and 30 mg TMA-N/lOOg in packed cod were rejected by sensory

panels (Dalgaard, Gram and Huss 1993), while 5mg TMA-N/lOOg has been the limit for

acceptable shrimp in some Australian and Japanese markets (Finne 1992). Nair and

Gopakumar (1986) reported that the levels of TMA-N of dehydrated salt mince threadfin

bream (Nenupterus japomcus) were in the range of 9.5-13 mg/lOOg, while sun dried and

dehydrated squid (Illex illecebrosus) were in the range of 5-94 mg/lOOg (Haard 1995).

Fresh Mod. spoiled Spoiled

Raw material quality

Fig. 4. 22 Trimethylamine nitrogen of dried salted fish

EH raw material ■dried salted product

Regarding raw material qualities, dried salted fish prepared from fresh and

moderately spoiled fish do not have any significant differences in TMA-N levels (p>0.05).

A significantly (p<0.05) higher TMA-N level was detected in dried salted fish prepared

from spoiled fish. This is parallel with the levels of TVB-N observed in this study. It seems

145

that spoilage of raw materials during the first 6h did not give any effect on the levels of both

TVB-N and TMA-N of the products.

It is believed that levels of TMA-N are correlated with the flavour characteristics of

fish and fish products. The correlation between TMA-N and flavour characteristics of

dried salted fish in this study will be discussed in Section 4.4.2.

4.4.2 Headspace analysis

As in the case for headspace analysis of sardine lipid in the previous section (Section

4.3), sniffing of GC effluents of the dried salted sardine was performed to determine the

odour characteristics of the individual aroma components, although further identification of

each volatile compound from the aromagram by GC-MS was not carried out. Additionally,

quantitative analysis was done based on relative proportions of the volatile compounds to

internal standard (undecane) that was added in each injected sample.

Different aroma characters resulting from the sniffing technique were detected in

this study, however, there were fewer chromatogram peaks than for sardine lipid. This

might be due to the fact that the major volatile compounds detected by GC originated

mainly from lipid. When dried salted sardines were analysed, their moisture contents were

much higher compared to those in the corresponding lipid. Moisture of samples might still

enter and accumulate inside the trap affecting the absorption of volatile compounds,

although detaching the trap from the DHS system and attaching it directly to the nitrogen

gas for a few minutes prior to chromatographic analysis was done to remove moisture from

the tenax.

Figure 4.23 shows the aromagram representing the odour character of individual

peaks. As for the odour of fish lipids (Section 4.3.2), the odour characteristics in this study

146

U(‘U

4

Axsra 'Xunq •

hd*o% ypo* k»9 e if *t/

jujnq Aysy pputj tli’rt

Aq*J ^po« Ai*> ttc

>»A

^sy >oo* A^p /f**c*

U&4 pouu u

rn

Ad«o« jogdum

*w

c»

jujnq ‘Aqsy A

jsnj i: i •?»

90cqqv3

p»foq »»r*

t

Xqsij jsp piouti

Xgsu Tujnq f*t**t^

^

I I/V

jCk

w-

j nrw

d6«gqv3MC

j etc *n

A*\j| *i» *e *

•••••I

Ajism A<TB)

Wo u

iw

Ml 'I

f»*

4*> »H »

flnqjfuq* m u

•«**?»u«oxr*nui M

J'tf>f,«—UXX)A/SnCLI

%*

A**y •*r-J*unq Aj*u

i *•»*♦!

Ar*»jfl

Ajruj 11 • •»I J2

3

***!•

T3a>

IC/}

Eoth*o<u£Ca.o.

a>cT3on

"O<u

7500

*o

_0J

pC3tbC3

r-~oro

CN

^r<uon

"u-

147

were also classified into six major groups according to Lin (1994), namely green, sweet,

oxidised, pesticide-like, cabbage-like and others and each class consists of minor notes as

shown in Table 4.9. In the early section of chromatogram (retention time between 2 to 4

min), the individual characteristic aromas could not be identified clearly since this area was

very complex and the peaks could not be separated well, therefore, for one large peak, the

mixture of aroma characters such as sweet, astringent, fruity, waxy and taint was detected

with unclear separation. Moreover, the chromatogram of this section was neither

reproducible, nor consistent. For this reason this area was not included for quantitative

analysis.

Table 4.9 The major and minor classification of aroma of dried salted sardines

Green Sweet Oxidised Pesticide- Cabbage- Otherslike like

fruity sweet fishy benzene rawcabbage

nutty

cucumber fried dried salted fish

cookedcabbage

meaty

stinkbug painty musty

burnt dirty sock

camphor

mushroom

malty

In this experiment green, oxidised and cabbage-like notes (Fig. 4.24) were detected

in higher level than other notes. Green odour has been recognised to be present in seafoods

both in freshly harvested as well as oxidised fish ( Josephson et al. 1983, 1987,

Karahadian and Lindsay 1989a) and the powerful plant-like odours are generally

148

contributed by unsaturated carbonyl and alcohol compounds (Kawai 1996). After

harvesting, enzymatic reactions occur as an initial oxidation of n-3 PUFA producing

unsaturated C6, C8 and C9 carbonyls and alcohols such as /ra«s-2-hexenal, m-3-hexen-l-ol,

cA-l,5-octadien-3-ol, /nms^-nonenal, frwzs,cA-2,6-nonadienal and l,5-octadien-3-one

contributing green cucumber, green stinkbug and plant-like odour (Karahadian and Lindsay

1989a). Non enzymic degradation of n-3 FA plays an important role in autoxidation with

hydroperoxides as the primary products (Lin 1994). These primary products lead to the

formation of carbonyl compounds with aldehydes as the major class of compounds

responsible for flavour reversion (Fujimoto 1989). The isomers of 2,4,7-decatrienals are

the typical secondary products in this autoxidation and trans,trans,cis-isomer contribute a

green-fishy character (Karahadian and Lindsay 1989a). For that reason, the green odours

detected in all dried salted sardines prepared from different stage of oxidised raw

materials and the more deteriorated the raw materials, the higher the quantity of green

notes (Fig. 4. 24).

The oxidised aroma including fishy, fried dried salted fish, painty and burnt were

present in high level (Fig. 4. 24) and were the dominant aroma observed in dried salted

sardines. These aromas were the characteristic notes of dried salted fish. Fishy off-flavour

has been recognised to occur in seafood and seafood products as a result of strongly

autoxidised PUFA of fish lipid producing saturated and unsaturated aldehydes (Meijboom

and Stroink 1972, Ke et al. 1975, Swoboda and Peers 1977, Karahadian and Lindsay

1989b). Among the aldehydes, trans,cis,cis-2,4,l-decatrienal is very important compound

responsible for fishy flavour (Meijboom and Stroink 1972, Badings 1973, Karahadian

and Lindsay 1989b). Figure 4. 25 illustrates the formation of 2,4,7-decatrienal as a result

of EPA decomposition during non-enzymic oxidation.

149

Con

cent

ratio

n (10

'6)8

7 Green

6

5

4

3

2

1

0

M **'is ■ V.:K®

■ a3

■• i

Sweet

i L

Cabbage-like

Fresh Mod. spoiled Spoiled Fresh Mod. spoiled Spoiled

':v: T

; ' -•r • ~

’JSprvgsai'

..

--- SHFresh Mod. spoiled Spoiled

Figure 4.24 Aroma characters of dried salted sardines prepared from different quality materials

149a

A=/WWWWv\Eicosapentaenolc acid

COOH

Toe opharol

/WWV^\RU c - hydroperoxide

/WWW-Ht^c,c-2,4(7-decatrienal

no Tocopherol

A4iWW*o9 ^ rotation (step 1)

•°0 /a A

11, translocation (step 2)

—A\OOH

i.i~ hydroperoxide

-A\ /„

JM, c- 2,4,7-decatrienal

Figure 4.25 The mechanism of the aldehyde formation in autoxidised EPA

(Karahadian and Lindsay 1989b)

Besides, due to aldehydes resulting from PUFA autoxidation, fishy aroma/flavour is

also contributed by TMA-N as a decomposition product of TMA-0 (Reineccius 1979,

Ashie et al. 1996, Kawai 1996). Different from the fishy aroma in sardine lipid that mainly

originated from lipid autoxidation, the fishy aroma in dried salted sardines was also caused

by decomposition of non-protein nitrogen compounds naturally present in fish muscle. The

reduction of TMA-0 is accomplished through bacterial enzymes during spoilage (Finne

1992, Ashie et al. 1996) therefore it is possible that the fishy aroma that occurred in dried

salted sardines was also contributed by TMA-N, since dried salted sardines were also

prepared from moderately spoiled and spoiled fish raw materials instead of the fresh one.

150

Moreover, TMA-N was also obtained in the products after salting and drying (dried salted

sardines) as illustrated in Fig. 4. 22. This was not surprising since heat enhanced the

decomposition of TMA-0 into TMA-N due to the instability of TMA-0 to heat (Colby et

al. 1993). This was in agreement with Japanese workers who found this amine in cooked

sardines (Koizumi et al. 1979) and dried salted sardines (Iida, Nakamura and Tokunaga

1979).

A combination of burnt, deep fried and metallic notes contributed to the overall

aroma that gave the impression of fried dried salted fish aroma. As noted by Karahadian

and Lindsay (1989b), distinct burnt /fishy aroma was caused principally by 2,4,7-decatrienal

isomers as a result of PUFA autoxidation, while /ra«s-2Tra/?s-4-decadienal, trans-2,trans-

4-nonadienal and /ra«5-2/ra«s-4-octadienal were reported to be responsible for deep fried

aroma (Saxby 1982, Triqui and Reineccius 1995a,b, Love 1996). Metallic aroma was

suggested to be associated with the fraction containing vinyl ketones such as oct-l-en-3-one

and octa-1, cA-5-dien-3-one resulting from n-6 PUFA and n-3 PUFA autoxidation,

respectively, by which the formation of these ketones was enhanced with the presence of

tocopherol-like antioxidants (Swoboda and Peers 1977, Karahadian and Lindsay 1989a). It

has been believed that fish, such as sardines, contain a significant amount of tocopherol

(Pigott and Tucker 1990, Izumi 1993), therefore the formation of such ketones resulting in

metallic notes during fish lipids autoxidation is understandable.

According to McGill et al. (1974), painty off-flavour in cold stored cod was

contributed by hept-4-enals, while Karahadian and Lindsay (1990) as well as Angelo et al.

(1988) suggested that hexanal, 2,4-heptadienals and 2,4-decadienals were responsible for

painty aroma in fish lipid. Short chain aldehydes such as /ra«s-but-2-enal and pent-2-enal

contributed to painty off odour in oxidised lipid (Saxby 1982, Hsieh et al. 1989). Painty

aroma detected in this study might be contributed by such aldehydes as in the case of heated

151

sardines that contain hexanal and hept-4-enals in the volatile compounds (Komi et al. 1992),

while Johnsen et al. (1992) noted that 2,4-heptadienals and 2,4-decadienals found in

frozen catfish during storage might contribute to painty off-flavour.

Regarding the overall oxidised fish aroma, dried salted sardine prepared from

spoiled raw material produced significantly higher level notes (p<0.05) than that prepared

from fresh and moderately spoiled raw materials. This indicates that the advanced

oxidation resulting in distinct oxidised off-flavour, especially fried dried salted fish aroma,

was found only in dried salted products from spoiled raw material.

The cabbage-like aroma was another significant volatile compound detected in this

study (Fig. 4. 24). It was reported that this aroma strongly related to sulphur-containing

compounds such as hydrogen sulphide (H2S), dimethyl sulphide (DMS), dimethyl disulphide

(DMDS), dimethyl trisulphide (DMTS) and methanethiol (Grosch 1982, Vejaphan et al.

1988, Milo and Grosch 1995). These volatile sulphur compounds were commonly

produced as a result of microbial activity, especially hydrogen-sulphide producing

organisms such as Pseudomonas sp and Alteromonas putrefaciens in decomposing seafood

samples (Lindsay, Josephson and Olafsdottir 1987, Ashie et al. 1996). Sulphur-containing

compounds were clearly obtained in cold stored fish, however, these compounds such as

H2S, DMS and methanethiol, were also determined in fresh salt-water fish (Kawai 1996).

This author commented further that the DMS present in shellfish and other saltwater fish

was derived from a precursor, dimethylpropiothetin (DMPT), contained originally in

microalgae, plankton and seaweed.

From the reason above, it can be assumed that the cabbage-like aroma detected in

this present study was contributed by such sulphur-containing compounds. However, it

could not be specified which compound was responsible for cabbage-like aroma since

further identification using GC-MS was not carried out. Moreover, the capability of the

152

FID GC system used in this study, in detecting sulphur compounds such as H2S was

limited.

Slightly different from the corresponding lipids, cabbage-like aroma was clearly

detected in the volatile compounds from dried salted sardines indicating that such aroma

might have originated from muscle rather than lipid. This is in agreement with Pippen et al.

(1969) who suggested that methionine was the possible precursor of cooked cabbage odour

in cooked poultry. Other workers found sulphur containing compounds such as DMDS and

DMTS in fresh oysters (Josephson, Lindsay and Stuiber 1985), blue-crab meat and blue

crab by processing (Chung and Cadwallader 1993), as well as in fermented fish and shrimp

(Cha and Cadwallader 1995); DMS and DMTS in heated sardine (Komi et al. 1992) and

DMDS in boiled and pasteurised crabmeat (Matiella and Hsieh 1990).

As in the case for oxidised aroma, the level of cabbage-like aroma in this study was

significantly higher (p<0.05) in dried salted sardine prepared from spoiled raw material than

in other products, while those in dried salted sardines prepared from moderately spoiled and

fresh raw materials were not different significantly (p>0.05).

Sweet, petroleum-like and other aroma including nutty, meaty, musty, dirty sock,

camphor, mushroom and malty detected in the aromagram were present in low level and

were not dominant in contributing aroma to dried salted sardines. There were no significant

differences (p>0.05) among these aroma detected in dried salted fish prepared from

different quality raw materials (Fig. 4. 24).

Sweet characters had been reported to be detected in seafoods (Josephson et al.

1987, Tanchotikul and Hsieh 1991, Cha et al. 1993). In refrigerated fish, a group of short

chain alcohols might contribute to sweet aroma and medium chain alcohols such as 3,6-

noadien-l-ol might impart heavy sweet such as sweet melon-like aroma character

(Josephson et al. 1987). Tanchotikul and Hsieh (1991) and Cha et al. (1993) reported that

153

ketones such as heptanone, cyclohexanone, 3-hydroxy-2-butanone and 6-methyl-5-hepten-

2-one were mainly responsible for sweet floral fruity notes in steamed rangia clam.

Josephson et al. (1987) observed the development of whitefish aroma during refrigerated

storage and they noted that the sweet stage was quite brief and immediately followed by a

stale phase.

Aromatic hydrocarbons originating from environmental pollution and/or

decomposition of carotenoid compounds might be responsible for the benzene-like aroma

detected in dried salted sardines. Reineccius (1991) suggested that alkylbenzenes

containing nine to ten carbons from petrochemical contamination might be responsible for

benzene or petroleum-like odours in seafoods. This supported other investigators who

noted that those off-flavours obtained in crustacean seafood and mullet caught in Australian

coastal waters (Vale et al. 1970, Whitfield and Freeman 1983) were from water pollution.

Carotenoids, one of the biological pigments that are commonly found in seafood

(Haard 1992), might also be present in sardines and might be responsible for the benzene­

like odour. During drying these compounds suffered from thermal decomposition

producing methyl naphthalenes, toluene and xylene, which imparted benzene-like odour

(Pippen et al. 1969). These compounds were detected in crustaceans (Vejaphan et al.

1988, Tanchotikul and Hsieh 1989, Matiella and Hsieh 1990) and in cooked chicken meat

(Nonaka et al. 1967) as well as in crude menhaden oils (Lin 1994).

Nutty and meaty aromas have been reported to be related to heterocyclic nitrogen

and/or sulphur-containing compounds that commonly contribute to desirable aroma in

crustaceans (Tanchotikul and Hsieh 1991, Cha et al. 1993, Cha and Cadwallader 1995).

Pyrazines are believed to contribute nutty, roasted and toasted characteristics to foods and

beverages and these compounds were also obtained in cooked crab (Chung and

Cadwallader 1993), shrimp paste (Cadwallader 1995), crayfish waste (Tanchotikul and

154

Hsieh 1989) and roasted peanut (Vercellotti et al. 1992). Pyrazines were reported to be

formed by Maillard reaction through Strecker degradation from various nitrogen sources

(Whitfield 1992). Interaction between unsaturated FA or their oxidation products with

sulphur-containing amino acids during heating produced the sulphur-containing

heterocyclic compounds such as thiazoles, thiophenes and dithiazines that were reported to

be important in generating meaty flavour in marine crustaceans (Tanchotikul and Hsieh

1989, 1991, Kawai et al. 1991, Cha etal. 1993).

Fish spoilage might produce degradation products such as H2S and ammonia. The

reaction of those compounds with aldehydes (resulting from enzymatic activity and/or

autoxidation) by the presence of heat during the drying process was the possible reason for

the formation of such heterocyclic compounds imparting nutty and meat aromas in this study.

Musty aroma detected in this present study might correlate with the spoilage process

during spoilage. According to Saxby (1982), the growth of Pseudomonas perolens in

chilled muscle produced volatile compounds including 2-methoxy-3-isopropylpyrazine that

was responsible for the musty odour, while Pippen et al. (1969) commented that isoleucine

and glutamic acid were the possible precursors for musty, rotten wood aromas in roasted

turkey. Blue-green algae producing geosmin seemed to be the dominant algae that caused

distinct musty flavour naturally occurring in pond-cultured fish (Lovell 1983, Johnsen et al.

1992).

Dirty socks aroma found in menhaden oil were reported to be contributed by fatty

acids such as butanoic, isobutanoic and hexanoic acids (Lin 1994). According to Grosch

(1982) acids might be formed by direct thermal decomposition of triacylglycerols,

hydrolysis of the ester linkages and autoxidation of aldehydes. Kawai (1996) proposed that

the presence of butanoic and pentanoic acids in grilled sardine and mackerel might be

formed from decomposition of salts of amines resulting from the removal of amines under

155

the reduced pressure. The dirty sock off-flavour detected in this study might be

contributed by low molecular weight volatile acids similar to that found in snow crab (Cha

et al. 1993), dried salted mullet and stravida (Fahmi, Soliman and Osman 1983), salted

donax and oyster (Soliman et al. 1983) and cooked sardine (Kawai 1996).

Undesirable camphor-like or naphthalene-like notes were detected in the aromagram

of dried salted sardines in this study. Pippen et al. (1969) reported that naphthalene odour

detected in cooked chicken meat related to naphthalene compounds, while Chung and

Cadwallader (1993) noted that indole has been described as having an undesirable

naphthalene-like odour. Naphthalenes were reported to be present in crayfish waste and

meat, refrigerated whitefish as well as in fermented fish and shrimp (Josephson et al. 1987,

Vejaphan et al. 1988, Tanchotikul and Hsieh 1989, Cha and Cadwallader 1995), while

indoles were found in volatile compounds of deteriorated white fish and heated sardine

(Josephson et al. 1987, Kawai 1996). The aromatic compounds might have originated from

the environment (Motohiro 1983, Reineccius 1991) or from thermal decomposition of

carotenoids since carotenoids have been reported as the precursors of naphthalene, toluene,

xylene, as well as unusual hydrocarbons and benzene derivatives found in aroma fractions of

foods including fish (Pippen et al. 1969).

Mushroom-like aroma that appeared in the sniffing runs in this study was also

detected in crayfish waste, boiled cod, boiled trout and ripened anchovy (Tanchotikul and

Hsieh 1989, Milo and Grosch 1995, Triqui and Reineccius 1995a,b). Oct-l-en-3-one has

been reported as the compound mainly responsible for mushroom aroma (Swoboda and

Peers 1977, Saxby 1982, Milo and Grosch 1995), however, Josephson et al. (1987)

proposed that, besides such ketones, Cg alcohols such as oct-1 -en-3-ol also contributed to

mushroom-like aroma. This was supported by Tanchotikul and Hsieh (1989), Love (1996)

156

as well as Triqui and Reineccius (1995a,b). These compounds were likely to be

enzymatically derived from PUFA (Josephson et al. 1987).

The characteristic aroma of malt was detected in the GC aromagram in this study.

According to Milo and Grosch (1995), aldehydes such as 2-methylpropanal, 2-

methylbutanal and 3-methylbutanal seemed to be responsible for malty aroma present in

boiled cod and trout, while Vercellotti et al. (1992) noted that methylpropanal contributed

to fermented dairy odour perceived in roasted peanuts. The aldehydes that might have

originated from lipid oxidation also occurred in fermented fish and shrimp, fresh sardine as

well as heated sardine (Cha and Cadwallader 1995, Kawai 1996).

Regarding the total volatile compounds that were estimated only from the detected

odour in aromagrams, the lowest level was obtained in dried salted sardine prepared from

fresh raw material while that prepared from spoiled raw material possessed the highest level.

As shown in Fig. 4.24, the high levels of volatile compounds had green, oxidised and

cabbage-like notes, while other notes were negligible. The different level of green and

cabbage-like notes clearly occurred among dried salted sardines at which the aroma level

increased with the decrease in raw material quality used for dried salted sardines.

However, significant oxidised aroma was produced only from dried salted sardine prepared

from spoiled raw material. The chromatograms of volatiles produced by dried salted

sardines prepared from different quality of raw materials are illustrated in Figure 4. 26.

In relation to aroma impression, the dominant aroma that characterised dried salted

fish was oxidised aroma, although a significant level of volatile compounds was also

obtained for green and cabbage-like aroma. It seems that oxidised aroma masked the

green, cabbage-like and other minor aroma compounds.

157

Figure 4.26 Chrom

atogram of dried salted sardine prepared from

different quality of raw

materials

158

4.4.3 Sensory evaluation

Different from the common practice of salting that is carried out in saturated brine

for overnight or 24h, salting in this experiment was of only 12h duration in saturated brine.

This time was chosen since, in the preliminary trial, salting for 24h in saturated brine

resulted in very poor appearance of the product with visible salt crystals on the surface and

brittleness, while the inner part of the product was moist (Fig. 4. 27). Moreover, moisture

loss and salt uptake during salting of the same fish in saturated brine ceased between the

sixth and tenth hours (Poemomo 1985), thus prolonged salting would be ineffective.

The results of sensory evaluation showed that appearance, colour and texture of raw

dried salted fish prepared from different quality raw materials gave significant differences in

the response from panelists, while aroma did not (Fig. 4.28). Dried salted fish prepared

from fresh fish were liked most by the panels based on appearance, colour and texture,

followed by dried salted fish prepared from moderately spoiled fish, while spoiled fish gave

the most disliked response. However, it is likely that the dislike response on raw dried

salted fish from spoiled fish was more due to the burst belly, blotchy skin, and slightly brittle

texture rather than aroma. Panelists consisting of Indonesian people could not distinguish

the aroma of raw dried salted fish prepared from different quality raw material.

As shown in Figs. 4.29 and 4.30, the physical condition of raw materials with

different quality was clearly visible, however, after processing, the differences were almost

undetectable except dried salted products prepared from spoiled fish that had burst belly and

faded skin. This suggests that immersing fish in saturated brine facilitated washing off

blood and slime resulting from the spoilage process, thus improving the final appearance of

the products.

159

Sens

orie

s

Fig. 4.27 Dried sardines after 12h and 24h salting A = 12h of salting, B = 24h of salting

8 APPEARANCEto

£ COLOURc3o

TEXTURE

AROMA

123456789

Hedonic scores (1-9)

Figure 4.28 Organoleptic acceptance of raw dried salted fish prepared from different raw material quality

□Fresh HiMod. spoiled ■ Spoiled

160

Salting for 12h seems to be sufficient in preventing further spoilage and drying

immediately after salting ceased further spoilage that might occur during process of salting.

Moreover, salt might combine with the protoplasmic anions of the microbial cells resulting

in a toxic effect, altering the enzymatic system of the microbial cell and lowering

proteolytic activity, as well as lowering water activity of the flesh (Frazier 1977),

consequently, these inhibit further spoilage. Discoloration and oily appearance that

commonly occur in dried salted fatty fish such as sardine, were not visible in this study,

suggesting that the lipid content of raw materials in this study were so low that it did not

create discoloration on the surface of skin. The lipid content of dried salted sardines in

this study was much lower (approx. 1% wb) compared to that of the same fish used in

previous studies that were 8-12% wb (Lubis 1990, Sphonphong 1990).

Fig. 4. 29 Sardines with different freshness condition A = Spoiled (10-h incubation), B = Moderately spoiled (6-hincubation), C = Fresh

161

Fig. 4. 30 Dried salted sardines prepared from different quality of raw materials.A = dried salted fish from fresh raw material, B = dried salted fish from

moderately spoiled raw material, C = dried salted fish from spoiled raw material

Regarding fried dried salted fish, the panelists were not able to distinguish the aroma

and taste of dried salted fish prepared from fresh and moderately spoiled raw materials, but

different response were obtained for dried salted products from spoiled raw material (Fig.

4.31). This was different from raw dried salted products, whereby panelists could not

distinguish the differences in aroma of the products. This was probably because, in testing

the fried products, in addition to sniffing the products, the panelists also tasted the

products. Thus the differences of the aroma and taste was clearer. Also, although frying

was only done for short time (40 second at 180°C) it could be possible, that during frying,

complex reactions occurred involving interaction between frying medium and the

products. Oxidation of unsaturated fatty acids was clearly pronounced whereby

hydroperoxides formed were very unstable at elevated temperature and would tend to

decompose generating flavour compounds (Reineccius 1994). In addition to oxidation of

unsaturated fatty acids, hydrolysis of TAG might occur during frying, while pyrolysis might

162

not have occurred in this study since the temperature of frying was much lower than 200°C

to allow pyrolysis. However, further study is needed to confirm the impact of frying on

the flavour development of fried dried salted sardines.

Otn•G0

143O

C/iCJ

'C

aCJ

C/3

AROMA

TASTE

1 2 3 4 5 6 7 8 9

Hedonic scores (1-9)

Figure 4.31 Organoleptic acceptance of fried dried salted fish prepared from different raw material quality

□ Fresh ^Mod. spoiled WSpoiled

In this study, score 5 (meaning neither like nor dislike) was set as the border score

for acceptance, as suggested by Larmond (1970) for a hedonic scale. It was shown that,

although all products were still not rejected, dried salted sardines prepared from spoiled raw

material were less preferred by panelists (Fig. 4.32), while only 5 of 35 panelists noted it as

rancid products. This could be due to the low lipid content of the raw materials,

consequently, the impact of lipid oxidation both during spoilage acceleration and processing

on flavour development, including those resulting from rancidity, was not detected clearly.

Some panelists even commented that the aroma of the raw products was not typical of

dried salted fish. It is common that dried salted fish after processing lacks aroma typical

of dried salted fish, and rancidity flavour commonly develops during storage (Ismail 1990,

Lubis and Buckle 1990).

163

123456789

Hedonic scores (1-9)

Figure 4.32 Overall acceptance of dried salted fish prepared fromdifferent raw material quality

□ Fresh ®Mod. spoiled HSpoiled

Another factor that might influence the response of panelists is that Indonesian

people are used to the strong flavour of dried salted fish commonly purchased from the

markets and these products have usually been stored for periods of time at ambient

temperature. It is necessary to evaluate the flavour of dried salted fish during storage in

correlation to the acceptability and headspace characteristics.

4.4.4 Relationship between sensory assessment and objective tests data

Sensory evaluation on the preference of dried salted sardines gave positive

correlations to chemical tests at which the panelists agreed that dried salted sardines

prepared from spoiled raw material were the less preferable than others. From chemical

analyse$;such dried salted sardines contained higher TVB-N and TMA-N than others. The

higher level of TVB-N might result in a higher level of ammonia since TVB-N consists of

TMA, DMA, ammonia and other volatile basic nitrogen compounds (Huss 1995). Thus the

strong ammonia aroma might affect the preference of the products by panelists. In the case

of TMA-N, it has been documented that those compounds were associated with typical

164

fishy aroma, hence higher level of TMA might also influence the acceptance of the

products.

As mentioned before, the most disliked products were dried salted sardine prepared

from spoiled raw material, while the liking of the products prepared from fresh and

moderately spoiled was not significantly different (p>0.05). In relation to volatile

compounds from headspace sampling, it was obtained that the increase in the total amount

of volatile compounds was parallel with a decrease in the raw material quality used for

dried salted fish processing. However, in the case of volatile compounds exhibiting

oxidised aroma (mixture of fishy, fried dried salted fish, burnt and painty), this increased

significantly only in dried salted sardine prepared from spoiled raw material. This means

that the least preference observed in the product from spoiled raw material correlated well

with the highest level of volatile compound exhibiting oxidised aroma in the same product.

This suggests that masking of aromas such as green, cabbage-like, and other minor aroma

compounds by the oxidised aroma resulted in different impact between total amount of

volatile aroma and oxidised aroma on the preference of the products by panelists.

165

5 CONCLUSIONS AND RECOMMENDATIONS

5.1 Conclusions

Fish lipids differ significantly from other animal lipids, especially in the more highly

polyunsaturated fatty acids compositions, which are at much higher levels. Therefore, fish

lipids are not as stable as those from plant and other animal oils. This has led to the need

to have suitable sample preparation and extraction methods that will be used for fish lipid

analysis in the laboratory in this study so that accurate information about the lipid

composition is obtained. This study has shown that fish sample treatment and extraction

methods affected the amount and quality of the lipids extracted from fish. In the method of

Hara and Radin (1978), the higher the proportion of fish tissue to solvents (a mixture of

hexane and isopropanol), the higher the amount of lipids recovered, and the lower the

protein contaminant. For the ratios tested (1:10 to 1:50), 1:50 ratio produced the highest

amount of extracted lipid, however due to practicality, this ratio was not used for

subsequent studies, and the ratio of 1:20 was used instead and compared with the method

of Bligh and Dyer (1959).

Compared to the method developed by Bligh and Dyer (1959), the Hara and Radin

(1978) method extracted more lipid, which had less protein contaminant and polar lipid.

However, in the stability test, it was shown that the lipids extracted by the method of Hara

and Radin (1978) suffered more autoxidation than that extracted by the method of Bligh

and Dyer (1959). This was indicated by higher weight gain and fluorescence during

incubation at 60°C for 12d. The method of Bligh and Dyer was then used in the

subsequent studies.

In an investigation of the spoilage pattern of Australian sardines (,Sardmella

neopilchardus) stored at a simulated Indonesian climate, it was demonstrated that pH could

166

not be used as a spoilage indicator. Other parameters monitored, i.e. sensory

characteristics (by demerit point scoring) and TVB-N (mg%N) showed a good correlation

with incubation time and spoilage. The increase of demerit point scores against incubation

time was well represented by Y = 1.93X + 5.68 (r = 0.99), where Y is the demerit point

score and X is the incubation time (h), while the correlation between TVB-N and

incubation time was exponential (Y = 10.98e°07X, r = 0.97, where Y is the TVB-N and X is

the incubation time). Since the correlation between the demerit point score and TVB-N

(mg%N) was found to be linear (Y = 1.48X - 8.13, r = 0.95), it can be concluded that

demerit point scoring, which is faster and easier to measure, can be used to predict the

TVB-N and hence the spoilage. It was also concluded that 6h incubation was able to

produce fish quality similar to that commonly used in the commercial production of dried

salted fish in Indonesia, while an incubation time of lOh was chosen to produce poor

quality fish which showed slight to moderate signs of spoilage. The first was referred to as

moderately spoiled, while the latter was classed as to spoiled fish in the making of dried

salted fish in subsequent studies.

Salting and drying were found to affect the TVB-N and TMA-N levels. The levels of

TVB-N and TMA-N increased markedly after processing, however, the differences in

quality of fresh and moderately spoiled fish were not detected in the TVB-N and TMA-N

levels of the final products.

The fat of the fish was also affected by salting and drying. It was shown that,

regardless the quality of the fish, the amount of fat decreased after processing, while the

signs of oxidation were clearly detected. The level of fluorescence of the lipids increased

markedly, while the polar compounds of lipids, except that from spoiled raw materials,

also increased. No changes in SAFA and MUFA were observed, but a significant decrease

167

in PUFA was noted after processing. Consequently, a decrease in PI was also observed,

since PI is the ratio of PUFA and Ci6:o-

There were many compounds detected by the FID, but only 24 and 18 were

detected by the nose having sensory significance in the lipids of dried salted fish and dried

salted fish respectively. Five sensory categories, namely green, sweet, oxidised/rancid,

pesticide-like and others were determined in the lipids of dried salted, while six sensory

categories, namely green, sweet, oxidised, pesticide-like, cabbage-like and others were

observed in dried salted fish. The fresher the raw materials, the less the amounts of the

volatile compounds detected in dried salted fish and in the corresponding oils. In general,

the amount of the volatile amounts in dried salted fish was less than that of the oils. Green

(apple, cucumber, melon, grass and stink bug flavours) and oxidised (oily, fishy, deep

fried, painty and burnt flavours) character notes were present in the highest levels. In

addition to green and oxidised character notes, the other volatile components in dried

salted fish, which were present in higher levels, were those of cabbage-like notes. Volatile

compounds such as benzene, which might have originated from the environment, were also

detected in both dried salted fish and its oils.

In relation to aroma impression, the dominant aroma that characterised dried salted

fish was an oxidised aroma (oily, fishy, deep fried, painty and burnt flavours), and this

aroma was found only in significantly higher level in dried salted sardine prepared from

spoiled raw material. Although green and cabbage-like aromas were also present in

significant levels, it seems that these aromas were masked by the oxidised notes. Sensory

evaluation showed that the most disliked products were dried salted sardine prepared from

spoiled raw material. On the other hand, the liking of the products prepared from fresh and

moderately spoiled fish was not significanty different. It was possible that the response of

168

the panelists was affected by the volatile compounds exhibiting oxidised aroma in the same

products.

Although the differences in the appearance of the raw materials were clearly

visible, the differences were almost undetectable in the products except those from spoiled

fish that had burst belly and faded skin. Discoloration and oily appearance that commonly

occur in dried salted fatty fish, such as sardine, were not visible in this study. Dried salted

fish prepared from fresh fish were liked most by the panels based on appearance, colour

and texture followed by moderately spoiled, while spoiled fish obtained the most disliked

responses.

5.2 Recommendations

Considering that dried salted fish is commonly served in fried form , while in

addition that oxidation of unsaturated fatty acids, hydrolysis of TAG and pyrolysis might

occur during frying, further study is needed to confirm the effect of frying on the flavour

development of fried dried salted sardines.

It is also necessary to evaluate the development of flavour characteristics of dried

salted fish during storage in correlation to the acceptability and headspace analysis, since

the characteristic flavour of dried salted fish develops during storage.

169

6 REFERENCES

Ackman, R.G. (1967) Characteristics of the fatty acid composition and biochemistry of

some fresh-water fish oils and lipids in comparison with marine oils and lipids.

Comp. Biochem. Physiol. 22: 907-922.

Ackman, R.G. (1980) Fish lipids: Part I. Connell, J.J. (ed.) Advances in Fish Science and

Technology. Farnham, Surrey, England: Fishing News Books Ltd.; 86-99.

Ackman, R.G. (1989) Nutritional composition of fats in seafoods. Prog. Food Nurt. Sci.

13: 161-241.

Ackman, R.G. (1992) Application of gas-liquid chromatography to lipid separation and

analysis: Qualitative and quantitative analysis. Chow, C.K. (ed.) Fatty Acids in

Foods and Their Health Implications. New York: Marcel Dekker Inc.; 47-63.

Ackman, R.G. and Gunnlaugsdottir, H. (1991) Seafoods and fishery by produtcs. Natural

and unnatural environments for longer chain omega-3 fatty acids. Angelo, A.J. St.

(ed.) Lipid Oxidation in Food. Washington, D C. Am. Chem. Soc.; 208-230.

Addison, R.F., Ackman, R.G. and Hingley, J. (1969) Free fatty acids of herring oils

possible derivation from both phospholipids and triglycerides in fresh herring. J.

Fish. Res. BdCan. 36 (6): 1577-1583.

Aitken, A., Jason, A.C., Olley, J. and Payne, P R. (1967) Effect of drying, salting and high

temperatures on the nutritive value of dried cod. Fishing News Int. 6 (9): 42-43.

Aitken, A., Lees, A. and Smith, J.G.M. (1984) Measuring Fish Composition. Torry

Advisory Note No. 89. Aberdeen: Edinburgh Press; 1 lp.

Angelo, A.J. St., Dupuy, HP. and Flick, G.J., Jr. (1988) Investigation of fish oil quality by

direct capillary gas chromatography. J. Food Qua/. 10: 393-405.

Anonymous (1991) Small Pelagic Fish Utilization: Research Needs. World Bank

170

Technical Paper No. 153. Fisheries Series. Washington, D.C.: The World Bank,

lip.

AO AC (1990) Official Methods ofAnalysis, 15th ed. Helrich, K. (ed.) Washington, D.C.:

AOAC.

AOCS (1990) Official Methods and Recommended Practices (4th edn). Method Ce lb-89.

Chicago, IL.: American Oil Chemists Society.

Ariyani, F., Fawzya, Y.N., Sari, A., Nasran, S. and Setiabudi, E. (1993) Effect of raw

material handling on quality of dried-salted ray. Liang, O.B., Buchanan, A. and

Fardiaz, D. (eds.) Development of Food Science and Technology in Southeast Asia.

Bogor, Indonesia: IPB Press; 307-315.

Arganosa, G.C. and Flick, Jr., G.J. (1992) Off flavors in fish and shellfish. Charalambous,

G. (ed.) Off Flavors in Foods and Beverages. New York: Elsevier Sci. Publ.; 103-

127.

Asakawa, T. and Matsushita, S. (1980) A colorimetric microdetermination of peroxide

values utilising aluminium chloride as the catalyst. Lipids 14: 865-967.

Ashie, I.N.A., Smith, J.P. and Simpson, B.K. (1996) Spoilage and shelf-life extension of

fresh fish and shellfish. Crit. Rev. Food Sci. Nutr. 36 (1&2): 87-121.

Astawan, M., Wahyuni, M., Yamada, K., Tadokoro, T. and Maekawa, A. (1994a) Effect of

high salt content of Indonesian dried-salted fish on rats. J. Agiic. Food Chem 42:

2265-2269.

Astawan, M., Wahyuni, M., Yamada, K., Tadokoro, T. and Maekawa, A. (1994b) Changes

in protein-nutritional quality of Indonesian dried salted fish after storage. J. Sci.

Food Agric. 66: 155-161.

Awad, A., Powrie, W.D. and Fennema, O. (1969) Deterioration of fresh-water whitefish

muscle during frozen storage at -10°C. J. Food Sci. 34: 1-9.

171

Badings, H.T. (1973) Fishy off-flavours in autoxidized oils. J. Am. Oil Chem. Soc. 50(8):

334.

Baumgartner, W.A., Baker, N., Hill, V.A. and Wright, E.T. (1975) Novel interference in

thiobarbituric assay for lipid peroxidation. Lipids 10 (5): 309-311.

Berhimpon, S., Souness, R.A., Buckle, K.A. and Edwards, R.A. (1990) Salting and drying

of yellowtail (Trachurus mccullochi Nichols). Int. J. Food Sci. Technol. 25: 409-

419.

Berhimpon, S., Souness, R.A., Driscoll, R.H., Buckle, K.A. and Edwards, R.A. (1991)

Salting behavior of yellowtail {Trachurus mccullochi Nichols). J. Food Proc. Pres.

15: 101-114.

Bidlack, W.R. and Tappel, A.L. (1973) Fluorescent products of phospholipids during lipid

peroxidation. Lipids 8 (4): 203-207.

Bilinski, E. and Lau, Y.C. (1969) Lipolytic activity toward long-chain triglycerides in

lateral line muscle of rainbow trout (Salmo gairdneri) J. Fish. Res. Bd Can. 26 (7).

1857-1866.

Bimbo, A.P. (1990) Production of fish oil. Stansby, M.E. (ed.) Fish Oils in Nutrition.

New York: Van Nostrand Reinhold, 141-180.

Bligh, E.G. and Dyer, W.J. (1959) A rapid method of total lipid extraction and purification.

Can. J. Biochem. Physiol. 37 (8): 911-917

Bligh, E G. and Merritt, J.H. (1988) New methods to reduce postharvest fishery losses in

traditionally processed and fresh fish in less developed countries. Morrissey, M.T.

(ed.) Postharvest Fishery Losses. Proceedings International Workshop Held April,

12-16, 1987, at The Univ. of Rhode Island, Kingston, RI. Rhode Island: ICMRD;

98-107.

Bligh, E G. and Scott, M.A. (1966) Lipids of cod muscle and effect of frozen storage. J.

172

Fish. Res. Bd Canada 23 (7): 1025-1035.

Bligh, E G., Shaw, S.J. and Woyewoda, A.D. (1988) Effects of drying and smoking on

lipids of fish. Burt, J.R. (ed.) Fish Smoking and Drying. The Effect of Smoking and

Drying on the Nutritional Properties of Fish. London: Elsevier Appl. Sci. ; 41-52.

Bonnell, A.D. (1994) Quality Assurance in Seafood Processing: A Practical Guide. New

York: Chapman & Hall; 54-76.

Bostock, T.W., Walker, D.J. and Wood, C.D. (1987) Reduction of Losses in Cured Fish

in the Tropics - Guide for Extension Workers. London: Trop. Dev. Res. Ins. 47p.

Bosund, I. and Ganrot, B. (1969) Lipid hydrolysis in frozen baltic herring. J. Food Sci. 34:

13-18.

Boyko, A.L., Morgan, M.E. and Libbey, L.M. (1978) Porous polymer trapping for GC/MS

analysis of vegetable flavors. Charalambous, G. (ed.) Analysis of Foods and

Beverages. New York: Academic Press; 57-79.

Branch, A.C. and Vail, A.M.A. (1985) Bringing fish inspection into the computer age.

Food Technol. Aust. 37 (8): 352-355.

Bremner, H.A., Statham, J.A. and Sykes, S.J. (1985) Tropical species from North-West

shelf of Australia: Sensory assessment and acceptability of fish stored on ice.

Reilly, A. (ed.) Spoilage of Tropical Fish and Product Development. Proceeding

of a Symposium held in conjunction with the Sixth Session of the Indo-pacific

Fishery Commision Working Party on Fish Technology> and Marketing. RMIT,

Melbourne, Australia, 23-26 October 1984. Rome: FAO; 41-53.

Brokerhoff, H. and Jensen, R.G. (1974) Lipolytic Enzymes. New York: Academic Press ;

330p.

Brown, A.J., Roberts, D.C.K. and Truswell, A.S. (1989) Fatty acid composition of

Australian marine finfish: A review. Food Aust. 41(3): 655-666.

173

Buckholz, L.L.J., Withycombe, D A. and Daun, H. (1980) Application and characteristics

of polymer adsorption method used to analyse flavor volatiles from peanut. J.

Agric. Food Chem. 28: 760-765.

Burgess, G.H.O., Cutting, C.L., Lovem, J.A. and Waterman, J.J. (1965) Fish Handling and

Processing. Edinburgh: Her Majesty's Stationery Office; 346-356.

Burhanuddin, Hutomo, M., Martosewojo, S. and Moelyanto, R. (1984) Sumber Daya Ikan

Lemuru (Indonesian Oil Sardines Resources). Jakarta: Lembaga Oceanologi

Nasional, LEPI. 70p.

Buttkus, H. (1967) The reaction of myosin with malonaldehyde. J. Food Sci. 32: 432-434.

Buttkus, H. (1975) Fluorescent lipid autoxidation products check. J. Agric. Food Chem.

23 (4): 823-825.

Buttkus, H. and Bose, R.J. (1972) Amine-malonaldehyde condensation products and their

relative colour contribution to the thiobarbituric acid test. J. Am. Oil Chem. Soc. 49:

440-443.

Caldironi, H.A. and Bazan, N.G. (1982) Effect of antioxidant on malonaldehyde

production and fatty acid composition in pieces of bovine muscle and adipose

tissue stored fresh and frozen. J. Food Sci. 47: 1329-1332, 1337.

Castell, C.H., MacLean, J. and Moore, B. (1965) Rancidity in lean fish muscle. IV. Effect

of sodium chloride and other salts. J. Fish. Res. Bd Can. 22 (4): 929-944.

Castell, C.H., Moore, B.A., Jangaard, P.M. and Weal, W.E. (1966) Oxidative rancidity in

frozen stored cod fillets. J. Fish. Res. Bd Can. 23: 1385-1401.

Cha, Y.J. and Cadwallader, K.R (1995) Volatile components in salt-fermented fish and

shrimp pastes. J. Food Sci. 60 (1): 19-24.

Cha, Y.J., Cadwallader, K.R. and Baek, H.H. (1993) Volatile flavor components in snow

crab cooker effluent and effluent concentrate. J. Food Sci. 58 (3): 525-530.

174

Chan, K.M. and Decker, E.A. (1994) Endogenous skeletal muscle antioxidants. Crit. Rev.

FoodSci. Nutr. 34 (4): 403-426.

Chang, S.S., Vallese, F.M., Hwang, L.S., Hsieh, O.A.L. and Min, D.B.S. (1977)

Apparatus for the isolation of trace volatile constituents from foods. J. Agric. Food

Chem. 25: 450-455.

Chio, K.S. and Tappel, A.L. (1969) Synthesis and characterization of the fluorescent

products derived from malonaldehyde and amino acids. Biochemistry 8: 2821-

2827.

Christie, W.W. (1982) Lipid Analysis: Isolation, Separation, Identification and Structural

Analysis of Lipids. 2nd edition. New York: Pergamon Press.

Chung, H.Y. and Cadwallader, K.R. (1993) Volatile components in blue crab (Callinectes

sapidus) meat and processing by-product. J. Food Sci. 58 (6): 1203-1211.

Colby, J.W., Enriquez-Ibarra, L.G. and Flick, G.J. Jr. (1993) Shelf life of fish and

shellfish. Charalambous, G. (ed.) Shelf Life Studies of Foods and Beverages.

Chemical, Biological, Physical and Nutritional Aspects. London: Elsevier Sci.

Publ.; 85-143.

Connell, J.J. (1980) Control of Fish Quality. Farnham, Surrey, England: Fishing New

Books Ltd. 222p.

Constantin, E. and Schnell, A. (1990) Mass Spectrometry. Chichester, England: Ellis

Horwood Limited. 184p.

Craske, J.D. (1995) A technique for manual injection to capillary columns with avoidance

of sample discrimination. J. Am. Oil Chem. Soc.: 995-999.

Craske, J.D. and Bannon, C.D. (1988) Letter to the editor. J. Am. Oil Chem. Soc.: 1190-

1191.

Cronin, D A. (1982) Techniques of analysis of flavours: Chemical methods including

175

sample preparation. Morton, I D. and MacLeod, A.J. (eds.) Food Flavors - Part A.

Introduction. New York: Elsevier Appl. Sci.; 15-48.

Dahle, L.K., Hill, E G. and Holman, R.T. (1962) The thiobarbituric acid reaction and

autoxidation of polyunsaturated fatty acid methyl esters. Arch. Biochem. Biophys.

98: 253-261.

Dalgaard, P., Gram, L. and Huss, H.H. (1993) Spoilage and shelf life of cod fillets packed

in vacuum or modified atmospheres. Int. J. Food Microbiol. 19: 283-294.

Dambergs, N. (1959) Extractives of fish muscle. 2. Solvent-water ratio in extraction of fat

and water solubles. J. Fish. Res. BdCan. 16 (1): 63-71.

Davis, L., Goodwin, L., Smith, G. and Hole, M. (1993) Lipid oxidation in salted-dried

fish: II. The effect of temperature and light on the rate of oxidation of a fish oil. J.

Sci. FoodAgric. 62: 355-359.

Dawes, P.A. (1993) Development, technology, and utilization of capillary column for gas

chromatography. Baugh, P.J. (ed.) Gas Chromatography. A Practical Approach.

New York: Oxford University Press Inc.: 71-102.

de Koning, A.J. and Silk, M.H. (1963) The 2-thiobarbituric acid reagent for determination

of oxidative rancidity in fish oils. J. Am. Oil Chem. Soc. 40: 165-169.

de Koning, A.J., Evans, A.A., Heydenrych, C., Purcell, C.J.V. and Wessels, J.P.H. (1985)

A critical investigation of a number of different methods of lipid determination in

fish meal with particular emphasis on correction required in these determinations.

J. Sci. Food Agric. 36: 177-185.

de Koning, A.J., Milkovitch, S. and Mol, T.H. (1987) The origin of free fatty acids formed

in frozen cape hake mince (Merluccius capensis, Castelnau) during cold storage at -

18°C. J. Sci. Food Agric. 39: 79-84.

de Koning, A.J. and Mol, T.H. (1990) Rates of free fatty acid formation from phopholipids

176

and neutral lipids in frozen cape hake (Merluccius spp) mince at various

temperatures. J. Sci. FoodAgric. 50: 391-398.

de Man, J.M., Tie, F. and de Man, L. (1987) Formation of short chain volatile organic

acids in the automated AOM method. J. Am. Oil Chem. Soc. 64 (7): 993-996.

Dillard, C.J. and Tappel, A.L. (1971) Fluorescent products of lipid peroxidation of

mitochondria and microsomes. Lipids 6 (10): 715-721.

Dillard, C.J. and Tappel, A.L. (1973) Fluorescent products from reaction of peroxidising

polyunsaturated fatty acids with phosphatidyl ethanolamine and phenylalanine.

Lipids 8 (4): 183-189.

Doe, P. and Olley, J. (1990) Drying and dried fish products. Sikorski, Z.E. (ed.) Seafood:

Resources, Nutritional Composition and Preservation. Boca Raton, FI.: CRC Press

Inc.; 125-145.

Dressier, M. (1979) Extraction of trace amounts of organic compounds from water with

porous organic polymers. J. Chromatogr. 165:167-206.

Drozd, J. and Novak, J. (1979) Headspace gas analysis by gas chromatography. J.

Chromatogr. 165: 141-165.

Dugan, L.R. Jr. (1955) Stability and rancidity. J. Am. Oil Chem. Soc. 32: 605-609

Dyerberg, J. (1986) Linolenate-derived polyunsaturated fatty acids and prevention of

atherosclerosis. Nutr. Rev. 44 (4): 125-134.

Entenman, C. (1957) General procedures for separating lipid components of tissue.

Colowick, S.P. and Kaplan, N.O. (eds.) Methods in Enzymology. Vol. III. New

York: Academic Press Inc.; 299-317.

Erickson, M.C. (1993) Lipid extraction from channel catfish muscle: Comparison of

solvent systems. J. Food Sci. 58 (1): 71-74.

Estrada, M., Olympia, M., Mateo, R., Milla, A., de la Cruz, A. and Embuscado, A. (1985)

177

Reilly, A. (ed.) Spoilage of Tropical Fish and Product Development. Proceedings

of a Symposium held in conjunction with the Sixth Session of the Indo-Pacific

Fishery Commission Working Party on Fish Technology and Marketing. RMIT,

Melbourne, Australia, 23-26 October 1984. Rome: FAO; 133-145.

Evershed, R.P. (1993) Combined gas chromatography-mass spectrometry. Baugh, P.J.

(ed.) Gas Chromatography. A Practical Approach. New York: Oxford University

Press Inc.: 359-391.

Eyers, C.W. (1991 ̂ Comparison of methods used in evaluating the oxidative stability of

fats and oils. Sydney: Department of Food Science and Technology, The

University of New South Wales. BSc thesis.

FAO (1981) The Prevention of Losses in Cured Fish. Rome: Food and Agricultural

Organization of the United Nations. 87p.

FAO (1995) 1993 Fishery Statistics. Vol. 76: Catches and Landing. Rome: Food and

Agricultural Organization of the United Nations. 425p.

Fahmi, A.A., Soliman, M.M. and Osman, F. (1983) Effect of salting treatment on the

flavour of stravida and mullet fish. Agric. Biol. Chem. 47 (1): 111-113.

Fazal, A. A. and Srikar, L.N. (1987) Changes in flesh lipids of seer fish during frozen

storage. J. Food Sci. Technol. 24: 303-305.

Finne, G. (1992) Non-protein nitrogen compounds in fish and shellfish Flick, J Jr. and

Martin, R E. (eds.). Advances in Seafood Biochemistry. Composition and Quality.

Lancaster, Pennsylvania: Technomic. Publ. Co.; 393-401.

Flanagan, R.J. (1993) Gas chromatography in analytical toxicology: Principles and

practice. Baugh, P.J. (ed.) Gas Chromatography. A Practical Approach. New

York: Oxford University Press Inc.: 171-212.

Fletcher , B.L., Dillard, C.J. and Tappel, A.L. (1973) Measurement of fluorescent lipid

178

peroxidation products in biological systems and tissues. Anal. Biochem. 52: 1-9.

Flick, Jr., G.J., Hong, G.P. and Knobl, G.M. (1992) Lipid oxidation of seafood during

storage. Angelo, A.J. St. (ed.) Lipid Oxidation in Food. Washington: Am. Chem.

Soc.; 184-207.

Folch, J., Ascoli, I., Lees, M., Meath, J.A. and Lebaron, F.N. (1951) Preparation of lipide

extracts from brain tissue. J. Biol. Chem. 191: 833-841.

Folch, J., Lees, M. and Stanley, G.H.S. (1957) A simple method for the isolation and

purification of total lipides from animal tissues. J. Biol. Chem. 226: 497-509.

Frank, J., Geil, J.V. and Freaso, R. (1982) Automatic determination of oxidative stability

of oil and fatty products. Food Technol. 36(6): 71-76.

Frankel, E.N. (1983) Volatile lipid oxidation products. Prog. Lipid Res. 22: 1-31.

Frazier, W.C. (1977) Food Microbiology 2nd edition. New Delhi : McGraw-Hill Publ. Co.

Ltd. 537p.

Fujimoto, K. (1989) Flavour chemistry of fish oils. Min, D.B. and Smouse, T.H. (eds.)

Flavour Chemistry of Lipid Foods. Champaign: Am. Oil Chem. Soc.; 190-195.

Garcia-Mesa, J.A., Luque de Castoo, M.D. and Valcarcel, M. (1993) Factors affecting the

gravimetric determination of the oxidative stability of oils. J. Am. Oil. Chem. Soc.

70 (3): 245-247.

Gelman, A., Pasteur, R. and Rave, M. (1990) Quality changes and storage life of common

carp (Cyprinus carpio) at various storage temperatures. J. Sci. Food Agric. 52: 231 -

247.

German, J.B., Zhang, H. and Berger, R. (1991) Role of lipoxygenases in lipid oxidation in

foods. Angelo, A.J. St. (ed.) Lipid Oxidation in Food. Washington: Am. Chem.

Soc.; 74-91.

Geromel, E.J. and Montgomery, M.W. (1980) Lipase release from lysosomes of rainbow

179

trout (,Salmo gairdneri) muscle subjected to low temperatures. J. Food Sci. 45: 412-

415, 419.

Gorczyca, E., Sumner, J.L., Cohen, D. and Brady, P. (1985) Mesophilic fish spoilage.

Food Technol. Aust. 37 (1): 24-26.

Graham, J. (1982) Freezing. Aitken, A., Mackie, I.M., Merritt, J.H. and Windsor, M.L.

(eds.) Fish Handling and Processing. Edinburgh: TorryRes. Stat.; 28-41.

Grant, D.W. (1996) Capillary Gas Chromatography. Chichester, England: John Wiley &

Sons Ltd. 295p.

Gray, J.I. (1978) Measurement of lipid oxidation: A review. J. Am. Oil Chem. Soc. 55:

539-546.

Gray, J.I. and Monahan, F.J. (1992) Measurement of lipid oxidation in meat and meat

products. Trends Food Sci. Technol. 3: 315-319.

Grob, K. (1973) Organic substances in potable water and its precursor. Part 1. Methods

for their determination by gas-liquid chromatography. J. Chromatogr. 84: 255-273.

Grosch, W. (1982) Lipid degradation products and flavour. Morton, I.D. and Macleod,

A.J. (eds.) Foods Flavours. Part A. Introduction. Amsterdam: Elsevier Sci. Publ.

Co.; 325-398.

Grundy, S .M. (1986) Effects of fatty acids on lipoprotein metabolism in man: Perspectives

for actions of fish oil fatty acids. Simopoulos, A.P., Kifer, R.R. and Martin, R E.

(eds.) Health Effects of Polyunsaturated Fatty Acids in Seafoods. New York:

Academic Press; 153-171.

Gudzinowicz, B.J., Gudzinowicz, M.J. and Martin, H.F. (1976) Fundamentals of

Integrated GC-MS. Part II: Mass Spectrometry. New York: Marcel Dekker Inc.

326p.

180

Haard, N. (1992) Biochemistry and chemistry of color and color change in seafoods.

Flick, G.J. and Martin, R E. (eds.) Advances in Seafood Biochemistry. Composition

and Quality. Lancaster, Pennsylvania: Technomic Publ. Co. Inc.; 305-360.

Haard, N.F. (1995) Chemical reactions and the quality of dried fish. Champ, B.R. and

Highley, E. (eds.). Fish Drying in Indonesia. Proceedings of an International

Workshop held at Jakarta, Indonesia on 9-10 February 1994. ACIAR Proceedings

No. 59; 67-75.

Hamberg, M., Niehaus, W.G. Jr. and Samuelsson, B. (1968) Preparation, isolation and

characterization of a derivative of malonaldehyde. Anal. Biochem. 22: 145-153.

Han, T.J. and Liston, J. (1987) Lipid peroxidation and phospholipid hydrolysis in fish

muscle microsomes and frozen fish. J. FoodSci. 52 (2): 294-296, 299.

Hansel, M., Rao, K.S., Matsuoka, T., Rali, T., Burrows, I. and Huber, M E. (1993) The

distribution of fatty acids in flesh and liver of Papua New Guinean Fish. Comp.

Biochem. Physiol. 106B (3): 655-658.

Hanson, S.W.F. and Olley, J. (1963) Application of the Bligh and Dyer method of lipid

extraction to tissue homogenates. Biochem. J. 89: 101-102.

Hara, A. and Radin, N.S. (1978) Lipid extraction of tissue with low-toxicity solvent. Anal.

Biochem. 90: 420-426.

Hardy, R. (1980) Fish lipids. Part 2. Connell, J. (ed.) Advances in Fish Science and

Technology. Farnham: Fishing News Books Ltd.; 103-111.

Hardy, R. McGill, A S. and Gunstone, F.D. (1979) Lipid and autoxidation changes in cold

storage cod (Gadus morhua) J. Sci. FoodAgric. 30: 999-1006.

Hicks, M. and Gebicki, J.M. (1979) A spectrophotometric method for determination of

lipid hydroperoxides. Anal. Biochem. 99: 249-253.

Hobbs, G. (1982) Changes in fish after catching. Aitken, A., Mackie, I.M., Merritt, J.H.

181

and Windsor, M.L. (eds.) Fish Handling and Processing. Edinburgh: Ministry of

Agric. Fish. & Food, Torry Res. Stat.; 20-27.

Hooyland, D.V. and Taylor, A.J. (1991) A review of the methodology of the 2-

thiobarbituric acid test. Food Chem. 40: 271-291.

Horner, W.F.A. (1992) Preservation of fish by curing (drying, salting and smoking). Hall,

G.M. (ed.) Fish Processing Technology. New York: VHC Publ.; 31-71.

Howgate, P.F. (1982) Quality assessment and quality control. Aitken, A., Mackie, I.M.,

Merritt, J.H. and Windsor, M L. (eds.) Fish Handling and Processing. Edinburgh:

Ministry of Agric. Fish. & Food, Torry Res. Stat.; 177-186.

Howgate, P.F. and Ahmed, S.F. (1972) Chemical and bacteriological changes in fish

muscle during heating and drying at 30°C. J. Sci. Food. Agric. 23: 615-627.

Hsieh, R.J. and Kinsella, J.E. (1989) Oxidation of polyunsaturated fatty acids: Mechanism,

products and inhibition with emphasis on fish. Adv. FoodNutr. Res. 33: 233-341.

Hsieh, T.C.Y., William, S.S., Vejaphan, W. and Meyers, S.P. (1989) Characterization of

volatile components of menhaden fish (Brevoortia tyrannus) oil. J. Am. Oil Chem.

Soc. 66(1): 114-117.

Hsieh, R.J., German, J.B. and Kinsella, J.E. (1988) Lipoxygenase in fish tissue: Some

properties of the 12-lipoxygenase from trout gill. J. Agric. Food Chem. 36: 680-

685.

Hultin, H O. (1991) Biochemical deterioration of fish muscle. Huss, H.H., Jakobsen, M.

and Liston, J. (eds.) Quality Assurance in the Fish Industry. Proceedings of an

International Conference, Copenhagen, Denmark, 26-30 August 1991. Amsterdam:

Elsevier Sci. Publ.; 125-138.

Hultin, H O. (1992a) Lipid oxidation in fish muscle. Flick, G.J. and Martin, R.E. (eds.)

Advances in Seafood Biochemistry. Composition and Quality. Lancaster,

182

Pennsylvania: Technomic Publ. Co. Inc.; 99-122.

Hultin, H O. (1992b) Trimethylamine-N-Oxide (TMA-NO) demethylation and protein

denaturation in fish muscle. Flick, J. Jr. and Martin, R E. (eds.). Advances in

Seafood Biochemistry. Composition and Quality. Lancaster, Pennsylvania: Techno.

Publ. Co. ; 25-42.

Huss, H.H. (1988) Fresh Fish - Quality and Quality Changes. A Training Manual. Rome:

FAO of the UN0.134p.

Huss, H.H. (1995) Quality and Quality Changes in Fresh Fish. FAO Fisheries Technical

Paper 348. Rome: FAO of the UNO. 195p.

Hwang, K.T and Regenstein, J.M. (1988) Protection of menhaden mince lipids from

rancidity during frozen storage. J. FoodSci. 54 (5): 1120-1124.

Iida, H., Nakamura, K. and Tokunaga, T. (1979) Changes in volatile components formed

in salted and dried sardine during frozen storage. Search for origin of volatile

components. Bull. Tokai Reg. Fish. Res. Lab. 54 (98): 77-85.

Ingenmansson, T., Kaufmann, P. and Ekstrand, B. (1993) Lipid hydrolysis and oxidation in

light and dark muscle of frozen stored rainbow trout (Oncorhynchus mykiss).

Ingemansson, T. (ed.) Lipid Composition and Deterioration in Light and Dark

Muscle of Rainbow Trout (Oncorhynchus mykiss) Goteborg: Chalmers Univ.

Technol. and The Swedish Inst. Food Res.; 79-104.

Ismail, N. (1990) Preservation of Marine Products by Salting and Drying. Sydney,

Australia: Department of Food Science and Technology, The University of New

South Wales, PhD Thesis; 397p.

Izumi, M. (1993) Nutrient content of selected marine products. SPC Fish. Newslett. 64:

25-36.

Jason, A C. (1965) Drying and dehydration. Borgstrom, G. (ed.) Fish as Food Vol. III.

183

Processing: Part I. New York: Academic Press; 1-54.

Jennings, W.G., Wohleb, R. and Lewis, M.J. (1972) Gas chromatographic analysis of

headspace volatiles of alcoholic beverages. J. Food Sci. 37: 69-71.

Jennings, W.G. and Filsoof, M. (1977) Comparison of sample preparation techniques for

gas chromatographic analysis. J. Agric. Food Chem. 25 (3): 440-445.

Johnsen, P.B., Dupuy, H.P., Legendre, M.G. and Flick, G.J. (1992) Instrumental analysis

of farm-raised catfish flavor quality. Flick, G.J and Martin, R.E. (eds.) Advances

in Seafood Biochemistry: Composition and Quality. Lancaster, P.A.: Technomic

Publ. Co. Inc.; 361-376.

Jones, N.R. (1967) Fish Flavours. Schultz, H.W., Day, E.A. and Libbey, L.M. (eds.) The

Chemistry and Physiology of Flavors. Westport, Connecticut: The Avi Publ. Co.

Inc.; 267-295.

Joseph, J.D. and Seaborn, G.T. (1982) Preliminary Studies in Marine Lipid Oxidation.

NOAA Technical Memorandum NMFS SEFC-95. Charleston, S.C.: US DC-

NOAA-NMFS. 92p.

Joseph, J.D. and Seaborn, G.T. (1990) The analysis of marine fatty acids. Stansby, M.E.

(ed.) Fish Oils in Nutrition. New York: Van Nostrand Reinhold, 40-72.

Josephson, D.B., Lindsay, R.C. and Stuiber, D A. (1983) Identification of compounds

characterizing the aroma of fresh whitefish (Coregonus clupeaformis) J. Agric.

Food Chem. 31: 326-330.

Josephson, D.J., Lindsay, R.C. and Stuiber, D.A. (1985) Volatile compounds

characterizing the aroma of fresh Atlantic and Pacific oysters. J. Food Sci. 50: 5-9.

Josephson, D.B., Lindsay, R.C. and Olafsdottir, G. (1987) Measurement of volatile aroma

constituents as a means for following sensory deterioration of fresh fish and fishery

products. Kramer, D.E. and Liston, J. (eds.) Seafood Quality Determination.

184

Proceedings of the International Symposium on Seafood Quality Determination

coordinated by the University of Alaska Sea Grant College Program, Anchorage,

Alaska, USA, 10-14 November 1986. New York: Elsevier Appl. Sci.; 27-47.

Junaeda, P. and Rocquelin, G. (1985) Rapid and convenient separation of phospholipids

and non phosphorus lipid from rat heart using silica cartridges. Lipids. 20: 40-41.

Kaitaranta, J.K. (1992) Control of lipid oxidation in fish oil with various antioxidative

compounds. J. Am. Oil Chem. Soc. 69 (8): 810-813.

Kaitaranta, J.K. and Ke, P.J. (1981) TLC-FID assessment of lipid oxidation as applied to

fish lipids rich in triglycerides. J. Am. Oil Chem. Soc. 58 (6): 710-713.

Kakuda, Y., Stanley, D.W. and Van de Voort, F.R. (1981) Determination of TBA number

by HPLC. J. Am. Oil Chem. Soc. 58: 773-775.

Kanner, J. and Kinsella, J.E. (1983) Lipid deterioration initiated by phagocytic cells in

muscle foods: J3-carotene destruction by a myeloperoxidase-hydrogen peroxide-

halide system. J. Agric. Food Chem. 31: 370-376.

Karahadian, C. and Lindsay, R.C. (1989a) Evaluation of compounds contributing

characteristic fishy flavors in fish oils. J. Am. Oil Chem. Soc. 66 (7): 953-960.

Karahadian, C. and Lindsay, R.C. (1989b) Role of oxidative processes in the formation

and stability of fish flavors. Teranishi, R, Buttery, R.G. and Shahidi, F. (eds.)

Flavor Chemistry. Trends and Developments. ACS Symposium Series 388.

Washington, D C.: Am. Chem. Soc.; 60-75.

Karahadian, C. and Lindsay, R.C. (1990) Low temperature deodorizations of fish oils with

volatile acidic and basic steam sources. J. Am. Oil Chem. Soc. 67 (2): 85-91.

Kates, M. (1986) Techniques of Lipidology: Isolation, Analysis and Identification of

Lipids. 2nd edition. Amsterdam: Elsevier Sci. Publ.

Kawai, T. (1996) Fish flavor. Crit. Rev. Food Sci. Nutr. 36 (3): 257-298.

185

Kawai, T., Ishida, Y., Kakiuchi, H., Ikeda, N., Higashida, T. and Nakamura, S. (1991)

Flavor components of dried squid. J. Agric. Food Chem. 39: 770-777.

Ke, P.J. and Ackman, R.G. (1976) Metal-catalysed oxidation in mackerel skin and meat

lipids. J. Am. Oil Chem. Soc. 53 (10): 636-640.

Ke, P.J., Ackman, R.G. and Linke, B.A. (1975) Autoxidation of polyunsaturated fatty

compounds in mackerel oil: Formation of 2,4,7-decatrienals. J. Am. Oil Chem. Soc.

52: 349-353.

Ke, P.J., Ackman, R.G., Linke, B.A. and Nash, D.M. (1977) Differential lipid oxidation in

various parts of frozen mackerel. J. Food Technol. 12: 37-47.

Ke, P.J., Cervantes, E. and Robles-Martines, C. (1984) Determination of thiobarbituric

acid reactive substances (TBARS) in fish tissue by an improved distillation-

spectrophotometric method. J. Sci. Food Agric. 35: 1248-1254.

Ke, P.J., Nash, D.M. and Ackman, R.G. (1977) Mackerel skin lipids as an unsaturated fat

model system for the determination of antioxidative potency of TBHQ and other

antioxidant compounds. J. Am. Oil Chem. Soc. 54: 417-420.

Keller, J.D. and Kinsella, J.E. (1973) Phospholipid changes and lipid oxidation during

cooking and frozen storage of raw ground beef. J. Food Sci. 38: 1200-1203.

Kelman, J.H. (1982) Flandling wet fish at sea. Aitken, A., Mackie, I.M., Merritt, J.H. and

Windsor, M.L. (eds.) Fish Handling and Processing. Edinburgh: Torry Res. Stat.;

28-40.

Khayat, A. and Schwall, D. (1983) Lipid oxidation in seafood. Food Technol. 37(7): HO­

MO.

King, M.F., Boyd, L.C. and Sheldon, B.W. (1992) Effect of phospholipids on lipid

oxidation of a salmon oil model system. J. Am. Oil Chem. Soc. 69 (3): 237-242.

Kinsella, J.E. (1987) Seafoods and Fish Oils in Human Health and Disease. New York:

186

Marcel Dekker, Inc.; 317p.

Kinsella, J. (1988) Fish and seafoods: Nutritional implications and quality issues. Food

Technol. 42 (5): 146-150.

Kinsella, J.E. (1990) Sources of omega-3 fatty acids in human diets. Lees, R.S. and Karel,

M. (eds.) Omega-3 Fatty Acids in Health and Disease. New York: Marcel Dekker,

Inc.; 157-199.

Kitson, F.G., Larsen, B.S. and McEwen, C.N (1996) Gas Chromatography and Mass

Spectrometry. New York: Academic Press 38lp.

Koizumi, C., Iiyama, S., Wada, S. and Nonaka, J. (1978) Lipid deterioration of freeze-

dried fish meats at different equilibrium relative humidities. Bull. Jap. Soc. Sci.

Fish. 44 (3): 209-216. Cited from the English version abstract.

Koizumi, C., Cao, thi K.T. and Nonaka, J. (1979) Undesirable odor of cooked sardine

meat. Bull. Jap. Soc. Sci. Fish. 45 (10): 1307-1312.

Kolarovic, L. and Fournier, N.C. (1986) A comparison of extraction methods for the

isolation of phospholipids from biological sources. Anal. Biochem. 156: 244-250.

Komi, H., Kawai, T. and Ishida, Y. (1992) Flavor components of heated sardine (in

Japanese) Abstracts of Nippon Suisan Gakkai, Spring Conference, Tokyo; 306.

Cited from Kawai (1996).

Krishnakumar, S., Hiremath, G.G and Menon, NR (1986) Preservation of Indian oil

sardine (Sardinella longiceps) in ice and chilled seawater. Part II. Changes during

storage with particular reference to salt penetration and lipid deterioration during

CSW holding. Fish. Technol. 23 (1): 1-16.

Krumperman, P.H. (1972) Erroneous peaks from porapak-Q traps. J. Agric. Food Chem.

20 (4): 909.

Krzynowek, J., Uljua, D.S., Panunzio, L.J. and Maney, R.S. (1992) Factors affecting fat,

187

cholesterol, and omega-3 fatty acids in Maine sardines. J. FoodSci. 57 (1): 63-65.

Lands, W.E.M. (1986) The fate of polyunsaturated fatty acids. Simopoulos, A.P., Kifer,

R.R. and Martin, R.E. (eds.) Health Effects of Polyunsaturated Fatty Acids in

Seafoods. New York: Academic Press; 33-48.

Larmond, E. (1970) Methods for Sensory Evaluation of Food. Canada: Canada

Department Agriculture; 57p.

Laubli, M.W. and Bruttel, P.A. (1986) Determination of the oxidative stability of fats and

oils: Comparison between the Active Oxygen Method (AOAC cd 12-57) and the

rancimat method. J. Am. Oil Chem. Soc. 63 (6): 792-795.

Lee, K.H., Jeong, I.H., Suh, J.S., Jung, W.J. and Kim, C.G. (1986) Utilisation of

polyunsaturated lipids in red muscled fishes. 1. Lipid composition of body oil and

lipids from different sections of sardine and mackerel. Bull. Korean Fish. Soc. 19

(5): 423-435.

Lewis, R.W. (1962) Temperature and pressure effects on the fatty acids of some marine

ectotherm. Comp. Biochem. Physiol. 6: 75-89.

Likens, S.T. and Nickerson, G.B. (1964) Detection of certain hop constituents in brewing

products. Am. Soc. Brew. Chem. Proc.; 5-13.

Lin, C.F. (1994) Flavor chemistry of fish oil. Ho, C.T. and Hartman, T.G. (eds.) Lipids in

Food Flavors. ACS Symposium Series, 558. Washington, DC.: Am. Chem. Soc.:

208-232.

Lindsay, R.C., Josephson, D.B. and Olafsdottir, G. (1987) Chemical and biochemical

indices for assessing the quality of fish packaged in controlled atmospheres.

Kramer, D.E. and Liston, J. (eds.) Seafood Quality Determination. Proceedings of

the International Symposium on Seafood Quality Determination coordinated by the

University of Alaska Sea Grant College Program, Anchorage, Alaska, USA, 10-14

188

November 1986. New York: Elsevier Appl. Sci.; 221-234.

Liston, J. (1980) Microbiology in fishery science. Connell, J.J. (ed.) Advances in Fish

Science and Technology. England: Fishing News Books, Ltd.; 138-157.

Lobb, K. (1992) Fatty acid classification and nomenclature. Chow, C.K. (ed.) Fatty Acids

in Foods and Their Health Implications. New York: Marcel Dekker, Inc.; 1-16.

Logani, M.K. and Davies, R.E. (1979) Lipid oxidation: Biologic effects and antioxidants -

A review. Lipids 15 (6): 485-495.

Love, J. (1996) Contributions of lipids to desirable and undesirable flavors in foods.

McDonald, R.E. and Min, D.B. (eds.) Food Lipids and Health. New York: Marcel

Dekker, Inc.; 287-314.

Love, R.M. (1982) Basic facts about fish. Aitken, A., Mackie, I., Merritt, J.H. and

Windsor, M L. (eds.) Fish Handling and Processing. 2nd edition. Aberdeen:

Torry Research Station; 2-19.

Love, R.M. (1994) Biochemical dynamics and the quality of fresh and frozen fish. Hall,

G.M. (ed.) Fish Processing Technology. New York: VCH Publ. Inc.; 1-30.

Lovell, R.T. (1983) Off-flavors in pond-cultured channel catfish. Water Sci. Technol. 15:

67-73.

Lovern, J.A. (1965) Some analytical problems in the analysis of fish and fish products. J.

Assoc. Offic. Agricl. Chem. 48 (1): 60-68.

Lovern, J.A. and Olley, J. (1962) Inhibition and promotion of post-mortem lipid hydrolysis

in the flesh offish. J. Food Sci. 27 (6): 551-559.

Lovern, J.A., Olley, J. and Watson, H.A. (1959) Changes in the lipids of cod during

storage in ice. J. Sci. FoodAgric. 10: 327-337.

Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) Protein measurement

with the Folin phenol reagent. J. Biol. Chem. 193: 265-275.

189

Lubis, Z. (1989) Studies on the Stability on Lipids in Dried Salted Sardines. Sydney,

NSW: Department of Food Science and Technology, The Univ. of New South

Wales. PhD Thesis. 302p.

Lubis, Z. and Buckle, K.A. (1990) Rancidity and lipid oxidation of dried-salted sardines.

Int. J. Food Sci. Technol. 25: 295-303.

Lundberg, W.O. (1967) General deterioration reactions. Stansby, M.E. (ed.) Fish Oils:

Their Chemistry, Technology, Stability, Nutritional Properties and Uses. London:

The Avi Pub. Co.; 141-147.

Mannan, B.A., Fraser, D.I. and Dyer, W.J. (1961a) Proximate composition of Canadian

Atlantic fish. I. Variation in composition of different sections of the flesh of

Atlantic halibut (Hippoglossus hippoglossus) J. Fish. Res. Bd Can. 18 (4): 483-

493.

Mannan, B.A., Fraser, D.I. and Dyer, W.J. (1961b) Proximate composition of Canadian

Atlantic fish. II. Mackerel, tuna and swordfish. J. Fish. Res. Bd Can. 18 (4): 483-

493.

Manurung, V.T. (1995) Salted fish consumption in Indonesia: Status and prospects.

Champ, B.R. and Highley, E. (eds.) Fish Drying in Indonesia. Proceedings of an

International Workshop held at Jakarta, Indonesia 9-10 February 1994. ACIAR

Proceedings (59), 25-30.

Marcuse, R. and Johansson, L. (1973) Studies on the TBA test for rancidity grading. II.

TBA reactivity of different aldehyde classes. J. Am. Oil Chem. Soc. 50: 387-391.

Markwell, M.A., Haas, S.M., Bieber, L.L. and Tolbert, N.E. (1978) A modification of the

Lowry procedure to simplify protein determination in membrane and lipoprotein

samples. Anal. Biochem. 87:211-222.

Maruf, F.W., Ledward, D.A., Neale, R.J. and Poulter, R.G. (1990) Chemical and

190

nutritional quality of Indonesian dried-salted mackerel (Rastrelliger kanagurta).

Int. J. Food Sci. Technol. 25: 66-77.

Matiella, J.E. and Hsieh, T.C.-Y. (1990) Analysis of crabmeat volatile compounds. J.

Food Sci. 55 (4): 962-966.

May, H.E. and McCay, P.B. (1968) Reduced triphosphopyridine nucleotide oxidase-

catalysed alterations of membrane phospholipids. I. Nature of the lipid alterations.

J. Biol. Chem. 243 (9): 2288-2295.

Maynard, J. (1983) Dried fish: An Asian staple food. INFOFISH Marketing Digest. (2):

18-22.

McGill, A S., Hardy, R. and Burt, J.R. (1974) Hept-cTM-enal and its contribution to the

off-flavour in cold stored cod. J. Sci. FoodAgric. 25: 1477-1489.

Meijboom, P.W. and Stroink, J.B.A. (1972) 2-Trans, 4-cis, 7-c/s-decatrienal, the fishy off-

flavor occurring in strongly autoxidized oils containing linolenic acid or w-3,6,9,

etc fatty acids. J. Am. Oil Chem. Soc. 49 (10): 555-558.

Melton, S.L. (1983) Methodology for following lipid oxidation in muscle foods. Food

Technol. 37 (7): 105-111,116.

Milo, C. and Grosch, W. (1995) Detection of odor defects in boiled cod and trout by gas

chromatography-olfactometry of headspace samples. J. Agric. Food Chem. 43:

459-462.

Min, D.B. and Kim, J.G. (1985) Gas chromatographic evaluation of flavour quality of oils.

Min, D.B. and Smouse, T.H. (eds.) Flavour Chemistry of Fats and Oils.

Champaign, 111.: Am. Oil Chem. Soc.; 241-262.

Miwa, K. and Ji., L.S. (1992) Laboratory Manual on Analytical Methods and Procedures

for Fish and Fish Products. 2nd. Singapore: Marine Fish. Res, Dep., SEAFDEC.

Motohiro, T. (1983) Tainted fish caused by petroleum compounds - A review. Water Sci.

191

Techno!. 15: 75-83.

Nambudiry, D.D. (1980) Lipid oxidation in fatty fish. The effect of salt content in the

meat. J. FoodSci. Technol. 17: 176-178.

Nair, PR. and Gopakumar, K. (1986) Development and storage characteristics of

dehydrated salt mince from low priced fish. Fish. Technol. 23: 101-105.

Nasran, S. and Arifudin, R. (1982) Study deterioratif lemuru pada suhu kamar dan suhu

dingin (Study on oil sardine deterioration at ambient and cool temperatures) (In

Indonesian) Nurhakim, S., Budihardjo and Supamo (eds.) Prosiding Seminar

Perikanan Lemuru (Proceedings of Seminar on Indonesian Oil Sardines Fisheries),

Banyuwangi, East Java, 18-21 January 1982. Jakarta, Indonesia: Puslitbang

Perikanan, Dept. Pertanian: 133-154.

Nasran, S., Ariyani, F., Fawzya, N. and Sari, A. (1992) Processing of dried salted ray at

Muara Angke. Proceedings of 8th session of the IPFC Working Party on Fish

Technology and Marketing, Yogya/carta, 24-27 September 1991. FAO Fish. Report

No. 470 Sppl.; 232-246.

Nelson, G.J. (1975) Isolation and purification of lipids from animal tissues. Perkin, E.G.

(ed.) Analysis of Lipids and Lipoproteins. Champaign, 111.: Am. Oil Chem. Soc ; 1-

22.

Nelson, G.J. (1991) Isolation and purification of lipids from biological matrices. Perkin,

EG. (ed) Analysis of Fats, Oils and Lipoproteins. Champaign, 111.: Am. Oil Chem.

Soc.; 21-60.

Nelson, J.S. (1994) Fishes of the World. 3rd ed. New York: John Willey & Sons, Inc.

600p.

Nichols, T.D., Nichols, D.S. and Volkman, J.K. (1993) Recent developments with marine

oil products in Australia. Chem. Aust. July: 33-40.

192

Nichols, T.D., Nichols, D.S. and Bakes, N.J. (1994) Marine oil products in Australia.

INFORM 5 (3): 254-261.

Nieto, S., Garrido, A., Sanhueza, J., Loyola, L.A., Morales, G., Leighton, F. and

Valenzuela, A. (1993) Flavonoids as stabilizers of fish oil: An alternative to

synthetic antioxidants. J. Am. Oil Chem. Soc. 70 (8): 773-778.

Nonaka, M., Black, D.R. and Pippen, E.L. (1967) Gas chromatographic and mass spectral

analysis of cooked chicken meat volatiles. J. Agric. Food Chem. 4: 713-717.

Ohkawa, H., Ohishi, N. and Yogi, K. (1978) Reaction of linoleic acid hydroperoxide with

thiobarbituric acid. J. Lipid Res. 19: 1053-1057.

Olafsdottir, G., Steinke, J.A. and Lindsay, R.C. (1985) Quantitative performance of a

simple tenax-GC adsorption method for use in the analysis of aroma volatiles. J.

FoodSci. 50 (5): 1431-1436.

Olcott, H.S. (1967) Antioxidants. Stanby, M.E. (ed.) Fish Oils: Their Chemistry,

Technology, Stability, Nutritional Properties and Uses. London: The Avi Pub. Co.,

Inc.; 164-170.

Olley, J., Doe, P.E. and Heruwati, S.E. (1988) The influence of drying and smoking on the

nutritional properties of fish: an introductory overview. Burt, J.R. (ed). Fish

Smoking and Drying. The Effect of Smoking and Drying on the Nutritional

Properties of Fish. New York: Elsevier Appl. Sci.; 1-21.

Olley, J. and Lovern, J.A. (1960) Phospholipid hydrolysis in cod flesh stored at various

temperatures. J. Sci. Food Agric. 11: 644-652.

Olley, J., Pirie, R. and Watson, H. (1962) Lipase and phospholipase activity in fish skeletal

muscle and its relationship to protein denaturation. J. Sci. Food Agric. 13: 501-516.

Opstvedt, J. (1988) Influence of drying and smoking on protein quality. Burt, J.R. (ed.)

Fish Smoking and Drying: The Effect of Smoking and Drying on the Nutritional

193

Properties of Fish. New York: Elsevier Appl. Sci.; 23-39.

Padley, F.B., Gunstone, F.D. and Harwood, J.L. (1986) Occurrence and characteristics of

oils and fats. The Lipid Handbook. Gunstone, F.D., Harwood, J.L. and Padley, F.B.

(eds.) London: Chapman and Hall; 49-141.

Palmer, F.B. St.C. (1971) The extraction of acidic phospholipids in organic solvent

mixtures containing water. Biochim. Biophys. Acta. 231:145-152.

Pedroza-Menabrito, A. and Regenstein, J.M. (1988) Shelf-life extension of fresh fish - A

review Part I - Spoilage of Fish. J. Food Qual. 11: 117-127.

Pedroza-Menabrito, A. and Regenstein, J.M. (1990) Shelf-life extension of fresh fish - A

review Part III - Fish quality and methods of assessment. J. Food Qual. 13: 209-

223.

Perez-Villarreal, B.P. and Howgate, P. (1991) Deterioration of European hake (Merluccius

merluccius) during frozen storage. J. Sci. FoodAgric. 55: 445-469.

Perkin, E G. (1975) Gas chromatography-mass spectrometry of lipids. Perkin, E.G. (ed.)

Analysis of Lipids and Lipoproteins. Champaign, Illinois: AOCS; 183-203.

Perkin, E.G. (1989) Gas chromatography and gas chromatography-mass spectropho­

tometry of odor/flavor components in lipid foods. Min, D.B. and Smouse, T.H.

(eds.) Flavor Chemistry of Lipid Foods. Champain, 111.: AOCS; 35-56.

Pigott, G.M. and Tucker, B.W. (1990) Seafood: Effects of Technology on Nutrition. New

York: Marcel Dekker, Inc. 362p.

Pikul, J., Leszczynski, D.E. and Kummerow, F.A. (1989) Evaluation of three modified

TBA methods for measuring lipid oxidation in chicken meat. J. Agric. Food Chem.

37: 1309-1313.t

Pippen, E.L., Mecchi, E.P. and Nonaka, M. (1969) Origin and nature of aroma in fat of

cooked poultry. J. Food Sci. 34: 436-442.

194

Pitt, J.I. (1995) Fungi from Indonesian dried fish. Champ, B.R. and Highley, E. (eds.).

Fish Drying in Indonesia. Proceedings of an International Workshop held at

Jakarta, Indonesia on 9-10 February 1994. ACIAR Proceedings No. 59; 89-96.

Poernomo, A. (1985) Salting and Drying Behaviour of Sardines. Sydney, Australia:

Department of Food Science and Technology, The University of New South Wales,

MAppSc thesis; 80p.

Poernomo, A. (1986) Drying behaviour of salted sardines. II. Effect of temperature and

its corresponding relative humidity. J. Post Harvest Fish. Res. 54: 47-53.

Poernomo, A., Authon, F. and Utomo, B.S.B. (1988) Salting and drying of fringescale

sardines {Sardinella fimbriata). II. Effect of initial water activity (aw) on product

quality. J. Post Harvest Fish. Res. 60: 17-25.

Polvi, S.M., Ackman, R.G., Lall, S.P. and Saunders, R.L. (1991) Stability of lipids and

omega-3 fatty acids during frozen storage of Atlantic salmon. J. Food Proc. Pres.

15: 167-181.

Prior, E. and Loliger, J. (1994) Spectrophotometric and chromatographic assays. Allen,

J.C. and Hamilton, R.J. (eds.) Rancidity in Foods. London: Chapman and Hall;

104-127.

Pryor, W.A., Stanley, J.P. and Blair, E. (1976) Autoxidation of polyunsaturated fatty

acids. II. A suggested mechanism for the formation of TBA-reactive materials from

prostalgandin-like endoperoxides. Lipids 11: 370-379.

Przybylski, R. and Eskin, N.A.M. (1995) Methods to measure volatile compounds and the

flavour significance of volatile compounds. Warner, K. and Eskin, N.A.M. (eds.)

Methods to Assess Quality and Stability of Oils and Fat-Containing Foods.

Champaign, 111.: AOCS Press; 107-133.

Putro, S. (1986) Better on-board handling of oil-sardines in the Bali Strait using chilled

195

seawater. INFOFISH Marketing Digest (1): 33-35.

Putro,S., Saleh, M. and Utomo, B.S.B. (1985) Storage life of rabbit fish (Siganus sp)

during icing. Reilly, A. (ed.) Spoilage of Tropical Fish and Product Development.

Proceedings of a Symposium held in conjunction with the Sixth Session of the Indo-

Pacific Fisheries Commission Working Party on Fish Technology and Marketing.

RMIT, Melbourne, Australia, 23-26 October 1984. Rome: FAO; 54-61.

Radin, N.S. (1981) Extraction of tissue lipids with a solvent of low toxicity. Lowenstein,

J.M. (ed.) Method in Enzymology, vol. 72. New York: Academic Press; 5-7.

Raghunath, M.R., Sankar, T.V., Ammu, K. and Devadasan, K. (1995) Biochemical and

nutritional changes in fish proteins during drying. J. Sci. Food Agric. 67: 197-204.

Ragnarsson, J O. and Labuza, T.P. (1977) Accelerated shelf-life testing for oxidative

rancidity in foods- A review. Birch, G.G and Green, L.F. (eds.) Food Chemistry 2.

London: Appl. Sci. Pub.: 291-308.

Rahardjo, S. and Sofos, J.N. (1993) Methodology for measuring malonaldehyde as a

product of lipid peroxidation in muscle tissues: A review. Meat Sci. 35: 145-169.

Rao, B.Y.K.and Bandyopadhyay, C. (1983) Lipid composition of salted sun-dried Indian

mackerel (Restrelliger kanagurta). J. Food Sci. Technol. 20: 62-64.

Reilly, A. and Santos, R.G. (1985) Histamine formation by bacteria isolated from bullet

mackerel (Auxis rochei Risso). Reilly, A. (ed.) Spoilage of Tropical Fish and

Product Development. Proceedings of a Symposium held in Conjuction with the

Sixth Session of the IPFC Working Party on Fish Technology and Marketing.

RMIT Melbourne, Australia, 23-26 October 1984. Rome: FAO; 359-362.

Reineccius, G. (1979) Off-flavors in meat and fish - review. J. Food Sci. 44 (1): 12-24.

Reineccius, G.A. (1985) Isolation, separation and characterization of flavour compounds in

lipids. Min, D.B. and Smouse, T.H. (eds.) Flavour Chemistry of Fats and Oils.

196

Washington: Am. Oil Chem. Soc.; 263-299.

Reineccius, G. (1989) Isolation of food flavors. Min, D.B. and Smouse, T.H. (eds.)

Flavour Chemistry of Lipid Foods. Champaign, IL. .Am.Oil Chem. Soc.; 26-33.

Reineccius, G. (1991) Off-flavors in foods. Crit. Rev. FoodSci. Nutr. 29 (6): 381-402.

Reineccius, G. (1994) Source Book of Flavours. 2 nd. ed. New York: Chapman & Hall;

24-60.

Reineccius, G. and Anandaraman, S. (1984) Analysis of volatile flavors. Lawrence, J.F.

(ed.) Food Constituents and Food Residues. Their Chromatographic

Determination. New York: Marcel Dekker, Inc. ; 195-293.

Reiss, U. and Tappel, A.L. (1973) Fluorescent product formation and changes in structure

of DNA reacted with peroxidising arachidonic acid. Lipids. 8(4): 199-202.

Roeraade, J. and Enzell, C.R. (1972) Tobacco chemistry: XIV. Sampling, concentration

and examination of tobacco headspace vapors. J. Agric. Food Chem. 20: 1035-

1039.

Rossell, J.B. (1989) Measurement of rancidity. Allen., J.C. and Hamilton, R.J. (eds.)

Rancidity in Foods. New York: Elsevier Appl. Sci.

Rutkowichz, S. (1982) Encyclopedia of Marine Fish. Wydawnictwo Morskie, Gdansk (in

Polish). Cited from Sikorski and Kamicki (1990).

Sahasrabudhe, M R. and Smallbone, B.W. (1983) Comparative evaluation of solvent

extraction methods for determination of neutral and polar lipids in beef. J. Am. Oil

Chem. Soc. 60 (4): 801-805.

Sakaki, T., Niino, K., Sakuma, H. and Sugawara, S. (1984) Analysis of tobacco headspace

volatiles using tenax-GC or active carbon. Agric. Biol. Chem. 48 (12): 3121-3128.

Saleh, M. and Murtini, J.T. (1982) Cara-cara pengolahan lemuru dan mutu hasil olahannya

di daerah Muncar (Processing method of sardines and their product quality in

197

Muncar) Nurhakim, S., Budihardjo and Suparno (eds.) Prosiding Seminar

Perikanan Lemuru (Proceeding of Seminar on Indonesian Oil sardines Fisheries)

Banyuwangi, East Java, 18-21 January 1982. Jakarta, Indonesia: Puslitbang

Perikanan, Dept. Pertanian; 155-162.

Sato, T., Matsuoka, N., Suhigara, H., Akazawa, H. and Motohiro, T. (1988) Petroleum­

like off flavor in seasoned herring roe. Water Sci. Technol. 20 (8/9): 49-53.

Sawicki, E., Stanley, T.W. and Johnson, H. (1963) Spectrophotometric and

spectrofluorometric methods for the determination of malonaldehyde. Anal.

Biochem. 35: 199-205.

Saxby, M.J. (1982) Taints and off flavors in foods. Morton, I.D. and Macleod, A.J. (eds.)

Foods Flavours. Part A. Introduction. Amsterdam: Elsevier Sci. Publ. Co.; 439-

457.

Schmedes, A. and Holmer, G. (1989) A new thiobarbituric acid (TBA) method for

determining free malonaldehyde (MDA) and hydroperoxides selectively as a

measure of lipid peroxidation. J. Am. Oil Chem. Soc. 66 (6): 813-817.

Schultz, T.H., Flath, R.A., Mon, T.R., Eggling, S.B. and Teranishi, R. (1977) Isolation of

volatile components from a model system. J. Agric. Food Chem. 25 (3): 446-449.

Shenderyuk, V.I. and Bykowski, P.J. (1990) Salting and marinating of fish. Sikorski, Z.E.

(ed.) Seafood: Resources, Nutritional Composition and Presentation. Boca Raton,

FI.: CRC Press Inc.; 147-162.

Sheppard, A.J., Hubbard, W.D. and Prosser, A.R. (1974) Evaluation of eight extraction

methods and their effects upon total fat and gas liquid chromatographic fatty acid

composition analysis of food products. J. Am. Oil Chem. Soc. 51: 416-418.

Shewan, J.M. (1955) The browning of salt cured white fish. FoodManuf. 30: 200-203.

Shewfelt, R.L. (1981) Fish muscle lipolysis - A review. J. FoodBichem. 5: 79-100.

198

Shewfelt, R.L., McDonald, RE. and Hultin, H O. (1981) Effect of phospholipid hydrolysis

on lipid oxidation in flounder muscle microsomes. J. FoodSci. 46: 1297-1301.

Shimasaki, H., Privett, O S. and Hara, I. (1977) Studies of the fluorescent products of lipid

oxidation in aqueous emulsion with glycine and on the surface of silica gel. J. Am.

Oil Chem. Soc. 54: 119-123.

Shono, T. and Toyomizu, M. (1972) Changes in fatty acids constituents of lipids in food

muscle during storage at low temperature (5°C). Decrease rate of C22.6 acid as a

criterion for the oxidative deterioration of lipid. Bull Jap. Soc. Sci. Fish 37 (9):

912.

Siang, N.C. and Kim, L.L. (1992) Determination of trimethylamine oxide (TMAO-N),

trimethylamine (TMA-N), total volatile basic nitrogen (TVB-N) by Conway’s

microdiffusion method. (1% boric acid and 0.02N hydrochloric acid). Miwa, K.

and Ji., L.S. (eds.) Laboratory Manual on Analytical Methods and Procedures for

Fish and Fish Products. 2nd. Singapore: Marine Fish. Res, Dep., SEAFDEC.

Sigurgisladottir, S. and Palmadottir, H. (1993) Fatty acid composition on thirty-five

Icelandic fish species. J. Am. Oil Chem. Soc. 70 (11): 1081-1087.

Sikorski, Z.E. (1990) Chilling of fresh fish. Sikorski, Z.E. (ed.) Seafood Resources,

Nutritional Composition, and Preservation. Boca Raton, FI.: CRC Press Inc.; 94-

109.

Sikorski, Z.E. and Karnicki, Z. (1990) Resources and their availability. Sikorski, Z.E. (ed.)

Seafood: Resources, Nutritional Composition and Preservation. Boca Raton: CRC

Presss, Inc.; 9-27.

Sikorski, Z.E. and Kolakowska, A. (1990) Freezing of marine food. Sikorski, Z.E. (ed.)

Seafood Resources, Nutritional Composition, and Preservation. Boca Raton, FI.:

CRC Press Inc.; 111-124.

199

Sikorski, Z.E., Kolakowska, A. and Burt, J.R. (1990a) Postharvest biochemical and

microbial changes. Sikorski, Z.E. (ed.) Seafood: Resources, Nutritional

Composition and Preservation. Boca Raton, FL: CRC Press Inc.; 55-75.

Sikorski, Z.E, Kolakowska, A. and Pan, B.S. (1990b) The nutritive composition of the

major groups of marine food organisms. Sikorski, Z.E. (ed.) Seafood: Resources,

Nutritional Composition, and Preservation. Boca Raton, FI.: CRC Press, Inc.; 29-

54.

Sikorski, Z.E. and Ruiter, A. (1994) Changes in proteins and non protein nitrogen in

cured, fermented and dried seafood. Sikorski, Z.E., Pan, B.S. and Shahidi, F.

Seafood Proteins. New York: Chapman & Hall; 113-126.

Sinclair, A.J. (1993) Nutritional significance of omega-3 polyunsaturated fatty acids for

humans. ASEAN Food Journal 8 (1): 3-13.

Sinclair, A.J., Dunstan, G.A., Naughton, J.M. Sanigorski, A.J. and O’Dea, K. (1992) The

lipid content and fatty acid composition of commercial marine and freshwater fish

and molluscs from temperate Australian waters. Aust. J. Nutr. Diet. 49 (3): 77-83.

Sinnhuber, R.O., Yu, T.C. and Chang, Y.T. (1958) Characterization of the red pigment

formed in the 2-thiobarbituric acid determination of oxidative rancidity. Food Res.

23: 626-633.

Sison, EC, Garcia, V.V., Carpio, E.V., Alcantara, P.R. Jr. and Madamba, C.S.P.( 1983)

Adaptation of artificial fish drying technology in the Philippines. James, D. (ed.)

The Production and Storage and Storage of Dried Fish. Universiti Pertanian

Malaysia, Serdang, Malaysia, 2-5 November 1982. FAO Fish. Rep., (279) Suppl.

Rome: FAO; 109-122.

Smith, G., Hole, G. and Hanson, S.W. (1990) Assessment of lipid oxidation in Indonesian

salted-dried marine catfish (Arms thalassinus) J. Sci. FoodAgric. 51: 193-205.

200

Soliman, M.M., Fahmy, A.A., El Sawy, A.A. and Osman, F. (1983) Effect of salting

treatment on the flavour of oyster, donax and sea urchin. Agric. Biol. Chem. 47(7):

1655-1659.

Souness, R. and Wibowo, S. (1994) Development of a rice-husk fired furnace and drying

chamber for fish drying in Indonesia. Champ, B.R. and Highley, E. (eds.) Fish

Drying in Indonesia. Proceedings of an International Workshop held at Jakarta,

Indonesia,9-10 February 1994. ACIAR Proceedings (59); 76-82.

Sophonphong, K. (1990) Effect of Modified Atmosphere Packaging on the Storage

Stability of Dried Salted Sardine (Sardinops neopilchardus). Sydney, Australia:

Department of Food Science and Technology, The University of New South Wales,

MAppSc Thesis; 134p.

Srikar, L.N., Seshadari, H.S. and Fazal, A.A. (1989) Changes in lipids and proteins of

marine catfish (Tachysurus dussumieri) during frozen storage. Int. J. Food Sci.

Technol. 24 (6): 653-658.

Stansby, M E. (1951) Report on fat in fish meal. J. Assoc. Offic. Agric. Chem. 34: 549-554.

Stansby, M.E. (1963) Industrial Fishery Technology. 2nd ed. New York: R.E. Krieger Pub.

Co. 415p.

Stansby, M E. (1971) Flavors and odors of fish oils. J. Am. Oil Chem. Soc. 48 (12): 820-

823.

Stansby, M E. (1982) Properties of fish oils and their application to handling of fish and to

nutritional and industrial use. Martin, R.E., Flick, G.J. and Ward D.R. (eds.)

Chemistry and Biochemistry of Marine Food Products. Westport, Connecticut:

AVI Publishing Company, Inc.; 75-92.

Stansby, M.E. (1990a) Introduction. Stansby, M.E. (ed.) Fish Oils in Nutrition. New

York: Van Nostrand Reinhold; 1-5.

201

Stansby, M E. (1990b) Deterioration. Stansby, M.E. (ed.) Fish Oils in Nutrition. New

York: Van Nostrand Reinhold; 120-140.

Sugiyono (1995) Problems associated with dried fish agribusiness in Indonesia. Champ,

B.R. and Highley, E. (eds.) Fish Drying in Indonesia. Proceeding of an

International Workshop held at Jakarta, Indonesia, 9-10 February 1994. ACIAR

Proceedings (59); 18-24.

Sumner, J.L. and Magno-Orejana, F. (1985) Do tropical fish keep longer in ice than

temperate fish. The circumstantial and definitive approaches. Reilly, A. (ed.)

Spoilage of Tropical Fish and Product Development. Proceed. Symp. Sixth Sess.

IPFC Work. Party, RMIT, Melbourne, Australia, 23-26 Oct. 1984. Rome: FAO

of UN; 62-70.

Suparno, Poemomo, A., Samianto, P., Setiabudi, E. and Subroto, W. (1989) Analysis on

Supply and Plant Site for Omega-3 Capsules (in Indonesian) Survey Report.

Unpublished.

Suzuki, T. (1981) Fish and Krill Protein: Processing Technology. London: Appl. Sci.

Publ. Ltd. 260p.

Suzuki, H. , Wada, S., Hayakawa, S. and Tamuro, S. (1985) Effect of oxygen absorber and

temperature on n-3 polyunsaturated fatty acids of sardine oil during storage. J.

Food Sci. 50: 358-360.

Swoboda, P A T. and Peers, K.E. (1977) Volatile odorous compounds responsible for

metallic, fishy taint formed in butterfat by selective oxidation. J. Sci. Food Agric.

28: 1010-1018.

Syamsiar, Fawzya, Y.N. and Poemomo, A. (1986) ACIAR Project 8313: Fish Drying in

East Java Area 3 : Salting and brining. The comparative characteristics of dry

salting and wet salting of oil sardines (Sardinella longiceps) J. Penelitian Pasca

202

Panen Perikanan. 56: 1-6.

Takiguchi, A. (1989) Effect ofNaCl on the oxidation and hydrolysis of lipids in salted

sardine fillets during storage. Bull. Jap. Soc. Sci. Fish. 55 (9): 1649-1654. Cited

from the English version abstract.

Tanchotikul, U. and Hsieh, T.C.-Y. (1989) Volatile flavor components in crayfish waste.

J. Food Sci. 54 (6): 1515-1520.

Tanchotikul, U. and Hsieh, T.C.-Y. (1991) Analysis of volatile flavor components in

steamed rangia clam by dynamic headspace sampling and simultaneous distillation

and extraction. J. Food Sci. 56 (2): 327-331.

Tarladgis, B.G., Watts, B.M., Younathan, M.T. and Dugan, L. Jr. (1960) A distillation

method for the quantitative determination of malonaldehyde in rancid foods. J.

Am. OilChem. Soc. 37: 44-48.

Tatum, V. and Chow, C.K. (1992) Effects of processing and storage on fatty acids in edible

oils. Chow, C.K. (ed.) Foods and Their Health Implications. New York: Marcel

Dekker, Inc.; 337-351.

Thorisson, S., Gunstone, F. and Hardy, R. (1992) The antioxidant properties of ethoxyquin

and of some of its oxidation products in fish oil and meal. J. Am. Oil Chem. Soc. 69

(8): 806-809.

Tipler, A. (1993) Gas chromatography instrumentation, operation, and experimental

considerations. Baugh, P.J. (ed.) Gas Chromatography. A Practical Approach.

New York: Oxford University Press Inc.; 15-70.

Triqui, R. and Reineccius, G.A. (1995a) Flavor development in the ripening of anchovy

(.Engraulis encrasicholus L.) J. Agric. Food Chem. 43: 453-458.

Triqui, R. and Reineccius, G.A. (1995b) Changes in flavor profiles with ripening of

anchovy {Engraulis encrasicholus). J. Agric. Food Chem. 43: 1883-1889.

203

Trombley, R. and Tappel, A.L. (1975) Fractionation and analysis of fluorescent products

of lipid peroxidation. Lipids. 10 (8): 441-447.

Turley, E. and Strain, J.J. (1993) Fish oils, eicosanoid biosynthesis and cardiovascular

disease: An overview. Int. J. FoodSci. andNutr. 44: 145-152.

Vale, G.L., Sidhu, G.S., Montgomery, W.A. and Johnson A.R. (1970) Studies on a

kerosene-like taint in mullet (Mugil cephalus) I. General nature of the taint. J. Sci.

FoodAgric. 21: 429-432.

van Klaveren, F.W. and Legendre, A. (1965) Salted cod. Borgstorm, G. (ed.) Fish as

Food. Vol. III. Processing: Part I. New York; Academic Press; 133-163.

van Veen, A.G. (1953) Fish preservation in Southeast Asia. Mark, E.G. and Stewart, G.F.

(eds.) Advances in Food Research. Vol. IV. New York: Academic Press.

Vejaphan, W., Hsieh, T.C.Y. and Williams, S.S. (1988) Volatile flavor components

from boiled crayfish (Procambarus clarkii) tail meat. J. Food Sci. 53 (6): 1666-

1670.

Vercellotti, R.J., Mills, O.E., Bett, K.L. and Sullen, D.L. (1992) Gas chromatographic

analyses of lipid oxidation volatiles in foods. Angelo, A.J.St. (ed.) Lipid Oxidation

in Food. ACS Symposium Series 500. Washington, D.C.: Am. Chem. Soc. ; 233-

265.

Walradt, J.P., Pittet, A O., Kinlin, T.E., Muralidhara, R. and Sanderson, A. (1971) Volatile

components of roasted peanuts. J. Agric. Food Chem. 19: 972-979.

Waterman, J.J. (1976) The Production of Dried Fish. FAO. Fish Technic. Paper (16):

52p.

Weber, P C., Fischer, S., von Schacky, C., Lorenz, R. and Stasser, T. (1986) Dietary

Omega-3 Polyunsaturated Fatty Acids and Eicosanoid Formation in Man.

Simopoulos, A.P., Kifer, R.R. and Martin, R.E. (eds.) Health Effects of

204

Polyunsaturated Fatty Acids in Seafoods. New York: Academic Press; 44-77

Wheaton, F.W. and Lawson, T.B. (1986) Processing Aquatic Food Products. New York:

A Wiley- Intersci. Publ. 327p.

Whitfield, F.B (1992) Volatiles from interactions of Maillard reactions and lipids. Crit.

Rev. FoodSci. Nutr. 31 (1): 1-58.

Whitfield, F.B. and Freeman, D.J. (1983) Off-flavors in crustaceans caught in Australian

coastal waters. Water Sci. Technol. 15: 85-95.

Wibowo, S., Sari, A., Suherman, M., Poemomo, A. and Souness, R. (1993) The

development of a fish dryer using rice husk furnace as heat source. Liang, O.B.,

Buchanan, A. and Fardiaz, D. (eds.) Development of Food Science and Technology

in Southeast Asia . Bogor, Indonesia: IPB Press; 184-188.

William, J.C., Field, R.A., Miller, G.J. and Welke, R.A. (1983) Evaluation of TBA

methods for determination of lipid oxidation in red meat from four species. J. Food

Sci. 48: 1776-1782.

Windsor, M and Barlow, S (1981) Introduction to Fishery By-Products. Farnham: Fishing

News Book Ltd. 187p.

Withycombe, D.A., Mookherjee, B.D. and Hruza, A. (1978) Isolation of trace volatile

constituents of hydrolyzed vegetable protein via porous polymer entrapment.

Charalambous, G. (ed.) Analysis of Foods and Beverages. New York: Academic

Press, 81-94.

Witte, V.C., Krause, G.F. and Bailey, M.E. (1979) A new extraction method for

determining 2-thiobarbituric acid values of pork and beef during storage. J. Food

Sci. 35: 582-585.

Wootton, M. and Ismail, N. (1986) The influence of processing conditions on the

properties of dried Nemadactylus macropetrus fillets. J. Sci. Food Agric. 37: 399-

205

408.

Yankah, V.V., Ohshima, T. and Koizumi, C. (1993) Effect of processing and storage on

some chemical characteristics and lipid composition of a Ghanaian fermented fish

product. J. Sci. FoodAgric. 63: 227-235.

Young, F.V.K. (1986) The chemical and physical properties of crude fish oils for refiner

and hydrogenators. Fish Oil Bull 18:1-19.

Yu, S.Y. (1995) Salted dried fish in Southeast Asia. Champ, B.R. and Highley, E. (eds.)

Fish Drying in Indonesia. Proceedings of an International Workshop held at

Jakarta, Indonesia 9-10 February 1994. ACIAR Proceedings (59); 44-50.

206

7. APPENDICES

7.1. Raw data and summary of results of statistical analysis for selection of extraction method for fish lipid

The following tables contain original data from which Figs. 4.1 - 4.5 have been drawn and their statistical data are presented in Table 7.1.4.

Table 7.1.1 Lipid yield extracted by the Hara and Radin (1978) method from different ratio of tissue to solvent (g/lOOg fish)

Tissue : solvent

No washing Washingrepl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

1:10 5.75 5.57 5.66 0.13 4.90 4.44 4.67 0.321:20 5.98 5.99 5.98 0.01 4.84 5.03 4.93 0.131:30 6.14 6.14 6.14 0.01 4.99 5.10 5.04 0.081:40 6.20 6.25 6.23 0.04 5.42 5.26 5.34 0.111:50 6.49 6.51 6.50 0.02 5.45 5.63 5.54 0.12

Table 7.1.2 Polar lipid of sardine lipid extracted by the Hara and Radin (1978) method from different ratio of tissue to solvent (g/lOOg lipid)

Tissue : No washing Washingsolvent repl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

1:10 22.82 23.60 23.21 0.55 14.46 14.68 14.57 0.161:20 22.58 22.61 22.60 0.02 14.29 14.20 14.25 0.061:30 21.98 21.71 21.85 0.19 13.54 13.89 13.72 0.251:40 21.98 21.27 21.63 0.50 13.58 13.72 13.65 0.101:50 21.38 21.72 21.55 0.24 13.45 13.84 13.65 0.28

Table 7.1.3 Protein content of sardine lipid extracted by the Hara and Radin (1978) method from different ratio of tissue to solvent (mg/g lipid)

Tissue : No washing With washingsolvent repl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

1:10 6.47 6.33 6.40 0.10 0.24 0.27 0.26 0.021:20 5.42 5.51 5.47 0.06 0.27 0.26 0.26 0.011:30 4.66 4.80 4.73 0.10 0.25 0.24 0.25 0.011:40 4.23 4.52 4.38 0.21 0.24 0.23 0.24 0.011:50 3.64 3.78 3.71 0.10 0.25 0.23 0.24 0.01

207

Table 7.1.4 Summary of results of statistical analysis for lipid extracted bythe Hara and Radin (1978) method from different ratio of tissue to solvent

tissue : Lipid yield (g/lOOg Polar lipid (g/lOOg Protein (mg/g lipid)fish) lipid)

solvent No washing Washing No washing Washing No washing Washing

1:10 5.66 a 4.67 a 23.21 a 14.57 a 6.40 a 0.26 a1:20 5.98 b 4.93 ab 22.60 ab 14.25 a 5.47 b 0.26 a1:30 6.14 be 5.04 ab 21.85 be U.llb 4.73 c 0.25 a1:40 6.23 c 5.34 be 21.63 c 13.65 b 4.38 d 0.24 a1:50 6.50 d 5.54 c 21.55 c 13.65 b 3.71 c 0.24 a

Values followed by the same letter in the columns within parameter (lipid yield, polar lipid and protein) are not significantly different (p>0.05)No correlation between data followed by the same letters for different columns

The following tables contain original data from which Figs. 4.6 - 4.7 have been drawn and their statistical data are presented in Tables 7.1.7 and 7.1.8.

Table 7.1.5 Weight gain of sardine lipids extracted by the Bligh and Dyer (1959) and the Hara and Radin (1978) methods during autoxidation at 60°C

Incubationtime

Weight change (%)

(days) Hara and Radin (1978) method Bligh and Dyer (1959) methodrepl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.001 0.37 0.39 0.38 0.01 0.02 0.01 0.02 0.012 0.86 0.83 0.85 0.02 0.09 0.10 0.09 0.003 1.44 1.39 1.41 0.03 0.18 0.17 0.17 0.004 1.92 1.83 1.87 0.06 0.26 0.24 0.25 0.015 2.39 2.23 2.31 0.11 0.35 0.37 0.36 0.016 2.78 2.59 2.69 0.14 0.43 0.45 0.44 0.027 3.09 2.89 2.99 0.14 0.53 0.56 0.54 0.028 3.39 3.18 3.29 0.15 0.68 0.77 0.72 0.069 3.69 3.53 3.61 0.11 0.90 1.02 0.96 0.0910 3.89 3.70 3.80 0.13 1.06 1.25 1.15 0.1311 4.10 3.97 4.03 0.09 1.33 1.58 1.45 0.1812 4.17 4.06 4.11 0.08 1.49 1.74 1.61 0.18

208

Table 7.1.6 Fluorescence of sardine lipids extracted by the Bligh and Dyer (1959) and the Hara and Radin (1978) methods during autoxidation at 60°C

Incubation Fluorescence (pg/g lipid)time Hara and Radin (1978) method Bligh and Dyer (1959) method

(days) repl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

0 0.23 0.18 0.21 0.03 0.31 0.30 0.31 0.012 5.83 5.49 5.66 0.24 4.18 3.93 4.05 0.184 7.87 7.24 7.56 0.44 6.49 6.38 6.44 0.086 8.27 8.14 8.20 0.09 7.16 6.68 6.92 0.348 8.75 8.58 8.67 0.12 7.22 6.98 7.10 0.1710 10.79 10.13 10.46 0.47 7.23 7.34 7.29 0.0812 11.62 10.85 11.23 0.54 7.98 8.55 8.27 0.40

Table 7.1.7 Summary of results of statistical analysis for lipid extracted by the Bligh and Dyer (1959) and the Hara and Radin (1978) methods during autoxidation at 60°C

Incubation time Weight gain(days) BD* HR**

0 0.00 a 0.00 a1 0.02 a 0.38 b2 0.09 ab 0.85 c3 0.17 ab 1.41 d4 0.25 be 1.87 e5 0.36 cd 2.31/6 0.44 d 2-69 g7 0.54 de 2.99 h8 0.72 e 3.29 i9 0.97/ 3.61 ij10 1.15* 3.80 j11 1.45 h 4.03 k12 1.61 i 4.11 /

* = Bligh and Dyer (1959) method ** = Hara and Radin (1978) method Values followed by the same letter in the same column are not significantly different (p>0.05)No correlation between data followed by the same letters for different columns

209

Table 7.1.8 Summary of results of statistical analysis for lipid extracted by the Bligh and Dyer (1959) and Hara and Radin (1978) methods during autoxidation at 60°C

Incubation time Fluorescence(days) BD* HR**

0 0.31 a 0.21 a2 4.05 b 5.66 b4 6.44 c 7.56 c6 6.92 cd 8.20 cd8 7.10 d 8.67 d10 7.29 d 10.46 c12 8.27 c 11.23 c

* = Bligh and Dyer (1959) method ** = Hara and Radin (1978) method Values followed by the same letter in the same column are not significantly different (p>0.05)No correlation between data followed by the same letters for different columns

7.2 Raw data and summary of results of statistical analysis for spoilage pattern of Australian sardine

The following tables contain original data from which Figs. 4.8, 4.9 and 4.11 have been drawn and their statistical data are presented in Table 7.2.4.

Table 7.2.1 Changes of demerit point score of sardine incubated at 28-30°C

Incubation Demerit point scoretime (h) repl.l repl. 2 repl. 3 aver. st.dev.

0 8 7 6 7 14 11 14 13 13 26 15 16 17 16 18 20 20 22 21 110 24 25 27 25 212 28 28 30 29 114 34 32 35 34 2

210

Table 7.2.2 Changes of TVB-N of sardine incubated at 28-30°C

Incubation TVB-N (mg% N)time (h) repl.l repl. 2 repl. 3 aver. st.dev.

0 11.54 11.72 11.91 11.72 0.24 15.88 15.27 15.35 15.50 0.36 16.60 16.71 16.82 16.71 0.18 17.28 17.73 17.68 17.56 0.210 20.42 20.70 20.86 20.66 0.212 22.82 22.45 23.59 22.95 0.614 30.30 29.53 31.06 30.30 0.8

Table 7.2.3 Changes of pH of sardine incubated at 28-30°C

Incubation pHtime (h) repl. 1 repl. 2 repl. 3 aver. st.dev.

0 6.1 6.1 6.1 6.06 0.04 6.2 6.1 6.0 6.10 0.16 6.2 6.2 6.2 6.19 0.08 6.2 6.1 6.1 6.12 0.010 6.1 6.1 6.1 6.08 0.012 6.2 6.0 6.2 6.11 0.114 6.2 6.2 6.2 6.24 0.0

Table 7.2.4 Summary of results of statistical analysis for demerit point score, TVB-N and pH of sardine incubated at 28-30°C

Incubation time (h)

Demerit point score

TVB-N pH

0 la 11.72 a 6.06 a4 13 b 15.50 b 6.10 ab6 16 c 16.71 c 6.19 b8 21 d 17.56 d 6.12 ab

10 25 6? 20.66 c 6.08 a12 29/ 22.95 / 6.11 ab14 34 g 30.30 g 6.24 c

Values followed by the same letter in the same column are not significantly different (p>0.05)No correlation between data followed by the same letters for different columns

211

Table 7.2.5 Correlation between time and demerit point, time and TVB-Nas well as demerit point and TVB-N of sardines incubated at 28-30°C

time vs demerit time vs TVB-N demerit vs TVB-Nx (time) y (dem) x (time) x (TVB-N) x (TVB-N) y (dem)

0 8 0 11.54 11.54 84 11 4 15.88 15.88 116 15 6 16.60 16.60 158 20 8 17.28 17.28 2010 24 10 20.42 20.42 2412 28 12 22.82 22.82 2814 34 14 30.30 30.30 340 7 0 11.72 11.72 74 14 4 15.27 15.27 146 16 6 16.71 16.71 168 20 8 17.73 17.73 2010 25 10 20.70 20.70 2512 28 12 22.45 22.45 2814 32 14 29.53 29.53 320 6 0 11.91 11.91 64 13 4 15.35 15.35 136 17 6 16.82 16.82 178 22 8 17.68 17.68 2210 27 10 20.86 20.86 2712 30 12 23.59 23.59 3014 35 14 31.06 31.06 35

0.99 correl. 0.94 correl 0.95 correl.5.68 intercept 10.23 intercept -8.13 intercept1.93 slope 1.18 slope 1.48 slope

y = 1,93x + 5,68________ y = 1.18x + 10.23______________y = 1.48x - 8,13

212

Table 7.2.6 Score sheet for sensory assessment

Sample code =

OBSERVATION SCORE OBSERVATION SCOREAPPEARANCE GILLS:very bright 0 colourbright 1 characteristic 0slightly dull 2 slightly dark, slightly faded 1dull 3 very dark, very faded 2

SKIN mucusfirm 0 absent 0soft 1 moderate 1

excessive 2SCALESfirm 0 smellslightly loose 1 fresh oily, metallic, seaweed 0loose 2 fishy 1

stale 2SLIME spoilt 3absent 0slightly slimy 1 BELLY:slimy 2 discolorationvery slimy 3 absent 0

detectable 1STIFFNESS moderate 2pre-rigor 0 excessive 3rigor 1post-rigor 2 firmness

firm 0EYES: soft 1Clarity burst 2clear 0slightly cloudy 1 VENT:cloudy 2 condition

normal 0Shape slightly break, exudes 1normal 0 excessive, opening 2slightly sunken 1sunken 2 smell

fresh 0Irish neutral 1visible 0 fishy 2not visible 1 spoilt 3

Blood BELLY CAVITY:no blood 0 stainsslightly bloody 1 opalescent 0very bloody 2 greyish 1

yellow-brown 2

bloodred 0dark red 1brown 2

MARKING very fresh = 0 very spoiled = 39

213

7.3 Raw data and summary of results of statistical analysis for lipid characterisation of raw material and their dried salted products

The following tables contain original data from which Figs. 4.12- 4.14 have been drawn and their statistical data are presented in Table 7.3.4.

Table 7.3.1 Lipid content (g/lOOg fish db) of raw sardine and the corresponding dried salted products

Raw mat. Raw material Dried salted productquality repl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

Fresh 3.96 3.93 3.95 0.02 1.79 1.79 1.79 0.00

Moderate- 3.75 3.57 3.66 0.13 2.05 1.97 2.01 0.06ly spoiled

Spoiled 4.00 3.75 3.88 0.18 2.15 1.90 2.03 0.18

Table 7.3.2 Fluorescence (pg/g fish) of raw sardine and the corresponding dried salted products

Raw mat. quality repl. 1

Raw material repl. 2 aver. st. dev.

Dried salted product repl. 1 repl. 2 aver. st. dev.

Fresh 0.020 0.020 0.020 0.00 0.213 0.207 0.210 0.00

Moderate- 0.024 0.023 0.024 0.00 0.243 0.230 0.235 0.01ly spoiled

Spoiled 0.023 0.023 0.023 0.00 0.253 0.229 0.240 0.02

Table 7.3.3 Polar compounds (g/lOOg lipid) of raw sardine and the corresponding dried salted products

Raw mat. Raw material Dried salted productquality repl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

Fresh 32.93 32.42 32.68 0.36 49.19 48.93 49.06 0.18

Moderate- 37.40 38.43 37.92 0.73 51.89 52.47 52.18 0.41ly spoiled

Spoiled 50.55 51.24 50.90 0.49 53.54 53.65 53.60 0.08

214

i3odOCOdu00

do'bdto

d 33 O3o m <L>

3

dT3I |3 |

^ aC/Dob o o<D h3d

3

3O .3

d

£

200c/3>%13Ccd13o V-» w '*—» i2C/34hOinr — ■*C/DoV-t4-i0b3s1300dco233H

£2<4-1Oind33OOh£ooMO

h

£*3

cr

“3vooCTvd

CO or-H VO2 2

id wo

3-3 0

CO (N O

VO o

\ C

\2 2

o

m ro *a

-O o

ooo CO VOm

il

o °

O

a <3

3OO OV Cvo o

o

© ©

o

-3-3

-0

<Ov i m

r-^ p

p

2 2 2

*3 3

3 WO VO oo ov vq oqC

O

CO

CO

d>,2

p2<gd3o24->

C

1 I

sgg

Oh<^

°o

<ulA

'S

3£32o300<3ti2<us

COG2<u£

'g<32d

jjo3 3

3 <«

•S «

^

+-“■3 2.a

>ooo

2

0 ^C

£ O

3% 2

^2

1 - 3

3O dO

h _

£

§

o <u

°

£

3 0

O ^2

3<S 23

3

o o

o o3a

o22

h.oo

jo 2d d

2

«3 |

M 22 2

t3 ii**

00oJ32>o£

oo2o£33Oh

Xod3o3OgoOh

d33ood3 W

3? §2

’B.

m2£

oo

O• i~H

H*_J

£,3

3? r/^3 3

33O'h-<004-H2OO<4-1o2

». < oo <u<41O

d2dbO3d3OOh

OOaboo<uddC* *3

<u3dSm*3<41o

**Oh

o o

o00

r+; r+; r-;Q

Hi Hi ^4

>-H04

*

c 2.3 a c 2.1 b c 2.1 b

* r t .CO —I VO

Mh00rH (N|

<Q

2 2

wo tN CS M

OhDOh3-0-0)

**

ov oo wo o oo t-

2 2

2CO CO CO

3 3

3O

hd d\ co

<

r t , C/0Qr- wo r^;2 2

2Mh52

♦3

3 3

*

<t (N t—oo vo d2 2

2

3 3

3*r t .

CO t"- COMhC/5

Cv CV O

<0hQ

OO OO C\

<003

3 3

Raw mat.quality RM5

Fresh 19.77od. spoiled 19.60 Spoiled 19.64

2

rH33O2.2

** = dried salted fishValues followed by the same letters in the columns within parameters (SAFA, MUFA, PUFA and PI) are n< No correlation between data followed by the same letters in the columns from different parameters

Table 7.3.6 Volatile compounds of sardine lipid extracted from different quality raw material (mg/kg)

odCOrdjj3aH3T3o3<Dc/3O&

OO AO AO t''- r-. O (N O

O CO

CO *—< AO N" ^

o

c?\ o ca

irit—

'O co oo

p

cq CN p CN

d

rd ao od cdt-"

ao co oo

OO CA •—I CA AO

CA O rd CA t "

AO t— CO

OO

oo ca r-—< ©

CN

AO CN OO CN cd ud

cq ao ca’—" CN cd

oo oo r^-CO CO t ■

(N tq oo co oi co co oo

<q p p

rq od od

cn od

co 1

oo ^r oo Caa© wd

cn ■'d r^ co

<—1

OO AO r-H oo Ca O

CN CNr- or

cn

CaCOr-CACNOKCNooCNCO

-t-Joo

o r~ <—i r^

r-I O OO ^

Oo

oo rq p

oo ao

n

od ca ao dAO

CO CN CO

»—; CN CA CO O0

n

ao rd od 6

AO CO CN

CO

AO CN CO AO t"-;d od od d

d

AO CO CO ■d"

n h- o

< cd CN

‘—J CO — oo

2 d

o ci

AO p OO

oo cd cnp AO oo

CA d ’—< AO

or cn r-

OO or CAca d

cdOO O

At ooCa ’—* ’—* CN OO CN

oo

r~~ or ’—i rd co '—■

' tq p cq

O r-

<—11 C

aor <n

ao

orodCOrdOOcd

o 03

• ’—>

o

II

AO OOo rd

«—1 oo

cd cd

p

rq rq p cq

cd rd od cd cn

AO ^ <—

11 CO

CN AO CA CA AO

cd •—* rd cn od

ao or -

a (N

o

n q

q o

of CN ca cd

Ca

AO OO —

' CO

OJ3

£

3 Cu O

' 3* 3 c3

O

3ooo

to3JOC/3

i*/c/3

oS2 ■ j§to

w

co r o

odd

rq or rd of

oo oo CO rd co -—

<

CA CA OOrd of

cn

«

73 -

o

d S

£ ©

§C/3 M-i

C/3

'oo£00

(N

CA CN

oo—: ©

© ’-d

oo co

r

oo© d

© od

CN '

co

ca r^ ca ao ci cd d

d

<—> y—

l CO

AO p p

—- -d

of

d

rd

CN T—l CO

>AJO>>.S

^©

n 3

O CXJO

8 ©• *—<

• »-H^

3 ^

* 3 .S2

AOdr~AOCOAO

216

Fresh raw material______ Moderately spoiled Spoiledrepl. 1 repl. 2 aver, st.dev. repl. 1 repl. 2 aver, st.dev. repl. 1 repl. 2 aver, st.dev.

Qh

sooJJrtj'o>VOcqr-do3Cd

H

c

o

•5 o c

o ^

" o c

--q CD

O

■'f; OO

of 0

—^0

00

0O in

cqo o o

OV

IT)

r-1 N

in

1/1 ----I

CN

co c

n •

on

vo cn

co i

vo

o

co

co v

od

dcd

cd

of^C

Nd

—<

cd co

cn —

1 cn

dd

dd

dd

iso

rid

of CN

1 ON

OM

Orni-

cN

iciO

Nn

d^ddrnd^dod

n't(N

01

'(N

h(N

'tco

d

oo

dcsd

^d

od

ri

^

<

co

cq '

CN oo

oocd

o

’—<

>—<

^

CN

(N

IT) h

; CO

O

'—<

CN CO

cO

t''-CN

CN

Oo

fOO

CN

CN

OO

OsO

CN

r^^O

O-^cd

oro

r-‘

mcnCO

CN O

3

orCO

C\

©—

< C

n CO

<—<

O O © O

iq

cq c

n c

n vq

of

vq—

oo o

CN o O

of —

' VO

CN t"' Cv

co oT of co

(N d 6 ^ d o

d

—I

CO

cq o

n cq

C5 oo

cn

of

cqo

r ©

on

»—<

cn

© © ©

of <

r"~

Gu,oJGo

3cd^

>N O

O

'O

o

v- 3

O

O

G

£*£ Z

’o.-S

Stt S

’0o

cd • **

cd G

<L> T

3 O

CO

cdVOcdof

of

VOCn

(Nofcqcoof

ofcovdCN

CNofcdCn

CNCNON

ON

CNCOCNOf

COCNOvdCOCNoovdVOCNrt3H

217

Table 7.3.7 Volatile compounds of lipid extracted from dried salted sardines prepared from different quality raw material (mg/kg)

O

to•S % S

g

£ 8

S

o 2

a

toco—

; n

o c

n —

- ^

oo K K

CN (N

(N

oo to in

to o o

CO

(N 6

'—<

*—<

CN

CO

oo

*-< O

O

CO

CO

CO

’—1

CN

CO

co vo

r-( oo cn

no

O

in ̂d o

coco

c\

co

o in

cotN

^

(N

CO

co to

to

cq co O

-I ini

»-H r-H ^

co t''- CN

O

Tf-co

cn c

n oo

C\ O

N

UO CO

O O

>-< O O C

\

co o

on

o

oocic

o'O

Om

i't CN

CN CN

Of

CN

^ ^

1-1 7

“q

cq

oo co

mi 6

6

cd CN

ON CN

■oj- OO

CN h; OO

CO O N

f O

n co

no

>—<

On <—

< CN

On

CN N

" O

CN

o r^- h

; m

o h-

co o o

-d o cd -d

q in

q cN

q

cdcN

oocN

r^'o

f —

i co

—<

co to

to

cq tq

on

r~- rq

co co

cn o

cor—

I oo

>—I

co o

q o q N

tq q

co CN

on CN fd cri

—I

OO >—

I CO to

£

o

g- o

ccS C

£3oCJo

obO

_ 3

g*7 CO

uco

05

Ca £ c5

) CO

CO

ON O <N

co d o

CN —

i o

mi ^ d

On co co

no

cq ON

oo >—

< o

nON

CO CN

O

*—1

co

h-

o

On

co co

CO —

1 7f

CN cd CN

to

fq CN

d d

^CN

CO

co

NO

On

NO

NO

CN

CN co

r- oo

toCN

NO ON CN

CN

CO

I C

Oo

dd

NO

t~" C

Oo

n -d

"d CN

7f

—1

ON 7f T

t r-l

<N

OO NO to

on—

I CN

O

S-. £3>

° §i/i

C w

cn o q q

cn oo

h h

cn o

cn ci

r—I

r—c CO

CN

O^qiqoq^f

cn

tri d

d

d n

on to

h

N

Nf (N

1 *—

c rj-

On

0O O

CO CO

dd

dd

dd

to CO

NO CN

NO O

f '

Of

—lO

OcN

CN

cq

-^t

d

rnn co d

d

iod

co oo

cn

-1 O

-7

1 "xf

^ O O q

N

cd cd

cd o

mi

Cn

O O

O rn

co

d d

d

cn d d

co

»—i

cn

oo

Tf

CN

O

OO

C

\ ■—1

d

d n

d

tcico

—

i c

n o

o

NO CN O

t"- CN

tod

'—1 d

CO

NO CO

CO

t-H

CN oo

to

cn

oo co

cq

oo

cd

cd

cd

cd o

-d

oo oo

td

rd

CO

NO t—

CN O

d

cn d

no

r- N

O

q

NO

■—'

"nf •—

1 fq

tri £

fd rd

rd -d co

-—

I N

O

< CN CN CN

(N cq

no d

oo cd d

dCO

’—I N

O

>7rC!

_co

73CC

O"33

~>

n-c

8\g

£

P3

co o

rt

33o

T3 a, jo

OZ00

t38 ^• f—t

• ^73

O*

9O cd

NOOOOCNCO

cqtj-ooC

OCNNOCNCOto7fNOcdtooooCOCNCN

NOCO

oNOCOr-CO(3coN£3Od)73£s jsP

i P

3

218

Tabic 7.3.7 Volatile compounds o f .......(continued)

| S3

si

oiio

'drl'o

dm

h6

c\

-—i m

1

»—1

—1

mo

ov

or'-

cv

in^m

^o

mcN

vd

cNC

T\o6

od-Nt;

cN’--<

oo cn cv

1 in co in ^

m

•—i «—

i m

CN

CN os m

'T t-* oo N"

OS Os CO •—

1 CN *-<

CN ;

OO COso

OS soos m

CO

N" <N

lO

OO

C

N rH

o

"t o ^

r-- ^

co >—

i —i

m

<—<

OO t"- T—11 os •*?od cc

d d

't

n

CO CN

CN VO CO —. so o

o o

o o

o o

oo in

co m

om

—ir^

-mm

ooc?N

'itor^

(N

dh

dm

m-ih

oo

dir)

t"-~ *—

< m

mm

N"CN

CN '—1O

OT

j-^O

Ot^

roO

C'\c

d'T

mm

»—

> vd co cv t"~ r-

. m

m N

- N

- cn

c^m

r'-cN

CN

Cv

cN

^to

oo

oddridddodi/od m ■—

1 m

m m

*—< N"CN

r-; ^O CO o

o< o

N; m

cn

cn cn

o —

^ m

co no

vo

mr-^

mcN

CN

-'d’

OC

OC

N

dd

'-irid

din

dd

d

m

t—i cn

co co r-

>—it-^

mN‘

r^cN

mN

"(N

co

h

in

^ri^m

tco

^l-

m

t—i cn

mm

co

oo o

mcn N

- m o

m m

m

cn m

m

t'-

"Hcd^ ^

*o x>rt

O ’O

coox5

>»CO3a

3 sT

3 O

cd • »—< cd

rdO TO

O

G

CNmCNmSOCNOmCNCNTfOmoomvooomoVON;ovoCNmcn

mmCNVOTf-

mmVON"OOVON"rt

H

219

Tabic 7.3.8 Summary of results of statistical analysis for volatile compounds of sardine lipid (x 10'5) extracted from raw sardines and their corresponding dried salted products

&a£

>

rt rt

q-

G

"G

r- co

vd<3

gtovo(NOCO

<5 G

CO

VOVO oC

N

'Cf

cn

n-

G

~G

Cn

CN C

N ^rd;

—I

CN

a

a -g

VO OO

CN ^

CN O

vd

to

cn

~g _

g

^3

^CN

oo g

VO

to

.fb ^ £

-G

Ga

a

<5h

to h

vq

r-

cn

O'—

* CN

*Ga

«gO

co

^

VO

tJ-

^vd

oo Q

!

g

-g

^00 ^

OO

to vo

. co

f**~ ’~~l

•G

G

~G

CO CN

O

^

; O

CN CO

CO

G

a

G

O

VO CO

CN CN

vo

G

*G

"G

VC OO

O'T

"Tto

CN

CO

GG

G O

co

VO tf

CN r-;

toC

O

CO

M

T3

= -o

o o

a

toocdVo&0gcdo01

o

-*—*

GcdSO§> c■

s a

<D0G!g1toGa3Ooo•G_G(3*

oc£

373ooGGtooo<DG3oGG

3O£JO

T3Q’rtto

T3

<U " u T

3IIGGOQ

cd "cd dCG

o o

>N

£^

o G

G

_oC3T) h o oo £

oG

220

7.4 Raw data and summary of results of statistical analysis for flavour characterisation of dried salted sardines

Table 7.4.1 Summary of results of statistical analysis for proximate of dried salted sardines

Raw mat. quality Moisture Protein Lipid Ash Salt a w

Fresh 45.51 a 35.59 a 1.79 a 33.66 a 27.20 a 0.11 aMod. spoiled 43.98 b 31.19 b 2.01 a 33.10 a 26.48 a 0.11 aSpoiled 46.31 c 33.90 c 2.03 a 34.37 b 27.28 a 0.11 a

Values followed by the same letter in the same column are not significantly different (p>0.05)No correlation between data followed by the same letters in different columns

The following Tables contain original data from which Figs. 4.21 and 4.22 have been drawn and their statistical data are presented in Table 7.4.4.

Table 7.4.2 The TVB-N (mg% N) of different quality raw sardines and the corresponding dried salted products

Raw mat. Raw material Dried salted productquality repl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

Fresh 11.12 11.12 11.12 0.00 38.68 38.68 38.68 0.00

Moderate- 15.94 15.95 15.95 0.01 38.65 38.46 38.56 0.13ly spoiled

Spoiled 19.95 20.83 20.39 0.62 40.24 39.96 40.10 0.20

221

Table 7.4.3 The TMA-N (mg% N) of different quality raw sardines and the corresponding dried salted products

Raw mat. Raw material Dried salted productquality repl. 1 repl. 2 aver. st. dev. repl. 1 repl. 2 aver. st. dev.

Fresh 0.33 0.29 0.31 0.03 1.95 1.96 1.96 0.01

Moderate- 0.49 0.44 0.47 0.04 2.11 1.99 2.05 0.08ly spoiled

Spoiled 0.73 0.73 0.73 0.00 2.34 2.33 2.34 0.01

Table 7.4.4 Summary of results of statistical analysis for TVB-N and TMA-N of dried salted sardines

Raw mat. TVB-N TMA-Nquality RM* DSF** RM* DSF**

Fresh 11.12 a 38.68 d 0.31 a 1.96 dMod. Spoiled 15.95 b 38.5 6 d 0.47 b 2.05 dSpoiled 20.39 c 40.10 c 0.73 c 2.34 c

Values followed by the same letter in the column within parameter (TVB-N and TMA-N) are not significantly different (p>0.05)No correlation between data followed by the same letters in the columns from different parameters

222

ing Tables contain original data from which Fig. 4.24 has been drawn and their statistical data is presented in Table 7.4.6.£01

toto5c3a£733cr

o'OaoTOac33*ocl

<L>3*5cdooT3OooT3

.at34-iOooTO33OC4aoojo~c3o>40Tf;(bjo30c3H2. 8ii

h O M 0\

o o

p o

o o

© o

(N

i/3 (N

C\

ON

tT

q

r-~ 46 o

o

no

O

C\

Mco ^

n h

46 o

o vb

r-^ —i oo 40

CN

-sf \C

OO

40 o’ ©

vd

CN 40

CN C

no o

—o o

o o

40 40

Cn

(N t O

of

O O

40

oj- <—

< O

—'

©

CO

tT

OO

40 O

O 40

t"~ q q q

CO

M

O"

40-4 o

o 46

VO

VO

oo

CO

0 0

)0

0

o o o o

r-- 40 vo q n \r

h;

q o

o

-4

CN ' *_ 40

O (Nj 43 h

of

C)

O

of

' O

n O

On

-—<

(N

Nf

of

O) o

of

o

to-O

3v s ^O

' 3 M

3

O

3

2

3 o

<-3

O

40

3opa

CNOOorCNoroCN©VOCNOroO©VOCNOooCN©orCNooj-ooro(N©)CN©VOCN©uo£42oo£00

40 —

I o o

vo 40

©

CN O

CO 0 0 0 0)0

)

or

On Os

CO VOo

cn vq

ro co

CN ©

CN O 40

VO OO CO

CO 1

VO CN OO

CO '—i

>b O

CN O

40

CO O

40 CO '

or

co 40 co vqCN

O CN O

40

r~~i r co co or

cn cq (N o o

© ©

o o

o

OO t-- or VO 40

vq cn cq cq coO

(N O

of

CO CO O

VO CN

40 CN <N

cq cq-b

O

CN O

of

CN CN oo vo oo

OO

C

O O

O

CO

CO

—;

o —

i o

of

t'-- cn

40 co

,—1

»—<

cq c

n cq

or o o

o o

o

or C

n co Cv 4

0OO

—< C

n c

n CNr—

I O

>—1

O O

f

CN O

O 40 CN VO

cq ' c-

co cv'

o

■ o

cb

vo

_

h —, r—

or

Cn

CN -q

CN 40

■—1

O CN O

of

4-ioo•a1-8

1 e s

cE

ais

GO

ta•H

’ '

OD

t>

* g

O cb

orcq©oocqooro40CO©40cqooqoCOqoVOcqoCNOoVOcqOcqOoocqO<u3oN3Ox>ora'o

'—1 (N COcq

cq cq

o o o

or

—'

40OO

CO

•—I

—I —

, co

VO

CN

Cn

CO

cq

cqCN

r-b cb

(N

Cn

•—<

VO CN

Cn

rb -h' (N

■—, co r~"

<—' o

o o

o

40 O

O CO

V

O

OO

40 b 6

CN

CO

40 OO

rq

rq

orb

od

r- o oo

40 O 40

b b ci

40 VO CNcq cq o

o o

o

VO

CN CO

CO

<b

ob

Cn

r- co

cq rb O

CN

CNcqVO

Cn

cq q

O

CN

S 23-,

30

<utoaC0

30

to 3

0rt30

o'rt ^3o

u ^ 'oa

8

Otocd33-3

t)

223

Table 7.4.5 Volatile compounds o f ....... (continued)

■| 8 Oc

oor—(r-Hom

roooo

O

O-—

'OO

OO

oooooooo

v/o

or^co

coocn

coO

OO

'—'O

OO

rb

—i o

C'v <—lun

CN

^t"

-0

>0 0'*

—'O

OO

cb

OC

MT

l'OrH

't^ON

(Nir—iin^o

ooo^’ooorn

O-tO

W^IN

MM

oo

-—''-^ooo^

—0

0C

50

00

00

—1-—' ON OO O

co cr> coco (N

O O

CN 'i ^

—i

O O

O <N O

O O

'T

c^'OO

'Om

intN’—'in

000

—^000

i h- in o (N

n(N

O(N

M^n(N

o

oo

rio

oo

xr

^roooor-'Csir'^v/o

Op

pp

OO

Op

OO

OO

OO

OO

(N N h

O CN ^

if) '—1 O

OO

CN

OO

O'T

i'tNO

'nts^

O'O

(N

(N

'O

'—«

CN *—

' m

(No

oo

rsio

oo

^r

oocor-—

(N^N

OO

m'-'^

O0

00

—^000^

0^-0

^

^ n

M ^

* rt

3 £ I

3 I

s p

o s

aB

aG

G

G"ooUOocK(NocKooocbpoG"pt"~prbppoC\cbcoVOlbOCNvd

tooXO

>11 *

•4—»

H

«+-coCOPGGOOeooc3O>00CO

13GrtoCO

1<+HOM

W)"tos.s

CO P

<L>£

aO P

P

2

S "O G

o5 '£

00 o

vppbjjxaH

<utoJg 3c3u

a <1 o

^ ^

oo O

CO o

b b

cb

Q

Q

-Q

uo co co ’—< -—< r~-b

'T

rb

«£Sa

a -or- co i/o p p

p

CN

CN

CO

a a

aVO O

oo p p

po o

o

3 3 0

co cn vo (N co co P ^

vb

3 3

3 C

O

VO ^

p p

p

O O

O

3 <5 3

t"- o

r- o

r- b b

d

Pjg8- -3J,

8 38

’B o

i-i s-«

D«

X X

oo

<—1

C3<2

oCO

u(U

o G

O13

o> £

224

lowed by the same letter in the same column are significantly different (p<0.05) ition between data followed by the same letters for different columns

ing Tabic contains original data from which Figs. 4.28, 4.31 and 4.32 have been drawn and their statistical data are presented in Table 7.4.8.£I

rtOs£«+-i0»O

crCD

1soG<L>CdO-oa*u.g'OcdcoG£T3C/5O434hO_CdbdbOCOGocop'To3cdH

o_G"3cdCOT3O■3CO

GO-uT3£cd

C4

O O

VO CO

NO

ON

o o

o o

—I O

OO CO^

cn rb

CV O

OO ©

cb 'd" cb ■’d"

<n o

r- vo b b

b m

O OO

Cd VO

—<

O

—;

O

© © © ©

^ io in

co co co co

in ^ in h;

m m

cb m

co in •d"

oo co co co co

OO <N VO ©

■1

>—<

Cd

•—<

O ©

O O

oo cs o

o in b ri ri rb

C\ C\ o

o

CN (N rb cb

oo in

<N (N

<N rb

O

O

G

a0,0

Q, o

CJi H

< O

H <

o3x

GGo3COGo‘CGgo’Ep

HG

oo o

o ©

co r-

co co

Cd

m

co cb

in oo

m

co

cd oo

p o

o o

OS

OS

Cd

cb

CV O

S(N cb

o\ OO cb cb

VO O

O (N

o ©

r- o

cb cb

c-cboocbcd£ £

O

co Vh

CdH

GGcoOVh

o01H,oocdr H3dO>OGj—

<

OoG3G,ooo<

225

rcpl. 1 repl. 2 aver, st.dev. repl. 1 repl. 2 aver, st.dev. repl. 1 repl. 2 aver, st.dev.

Table 7.4.8 Summary of results of statistical analysis for sensory data of raw and fried dried sal' og-3eS00O

*3 <3

<5 —

I vo

CT\ © VO

(N n n

cr

a q

O

CN CN

O

oo VO

ri oi ri

35 <3 -cs

m

n

r-- oo CO(N

(N rl

<3 a

a m cn m

m h

ooro m

m

Q -C5 -C5vo vo

co h-

c-i rn cn

3s -o

3>o vo o

CO ^

o<N

ro ^

a -O <0

r-^ r-l

ONOO ^

o

ri

cn 1

-

T3

3^ "O

rrt

0/~)ooAa,0143rto«3cfcOtg'S 43i

?o c3 a £J373oo0$*3co *—.o£aooo

jo©T3

O£_o

.2oooou£c3too£JO-ao£

226

Table 7.4.9 Sensory evaluation of dried salted fish

Please evaluate the appearance and taste of dried salted sardine and fillthe characteristics score by crossing (x) the column according to your evaluation.

Sample code 1 |

1. Raw fish

like intensely dislike intensely1 2 3 4 5 6 7 8 9

AppearanceColourTextureAroma

2. Fried fish

like intensely dislike intensely1 2 3 4 5 6 7 8 9

AromaTaste

3. Are the samples rancid ? yes ___If your answer is no, continue to question 4

no

If your answer is yes, please mark the degree of rancidity by marking the following column according to your evaluation

slightly rancid very rancid1 2 3 4 5 6 7 8 9

4. From the overall evaluation of both the raw and fried dried salted sardine (question 1 to 3), please mark the degree of your preference on the sample by crossing the following column according to your evaluation

like intensely dislike intensely1 2 3 4 5 6 7 8 9

5. What is your comment about the samples ?

227