impact behaviour of hybrid gfrp-concrete beam ... - UNSWorks
Effect of raw material quality on lipids and flavour ... - UNSWorks
-
Upload
khangminh22 -
Category
Documents
-
view
0 -
download
0
Transcript of Effect of raw material quality on lipids and flavour ... - UNSWorks
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
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-
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
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-
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 4.18 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