A. N U T R I T I V E A S P E C T S
I t is not possible to discuss the numerous analyses reporting changes that take p l a c e in the h e a t processing of various kinds of fish. Nor do analytical methods used in early studies permit too reliable conclusions.
Tarr, in Chapter 6 of V o l u m e I I , reported on the nutritive implications of such changes. A thorough m a p p i n g of the c h e m i c a l transformations during h e a t processing of herring has b e c o m e available largely during t h e 60's. A m o n g the p h e n o m e n a studied are the appearance of carbonyl compounds, changes in nitrogenous compounds, and the conversion of creatine ( H u g h e s , 1 9 5 9 , 1 9 6 0 , 1961a, b ) . However, no comprehensive over-all picture regarding fish in general of the c h e m i c a l effects of heat
ing is currently available b u t one will gradually e m e r g e from this and similar investigations around the world.
T h e amino acid composition of canned fish does not change ap
preciably with the m e t h o d of heat processing ( P r o c t o r and Lahiry, 1 9 5 6 ) . W i t h respect to manufacture of canned fish for low-salt diets, Murray et al., ( 1 9 5 4 ) listed the sodium and potassium contents of salmon and tuna canned without salt; such packs contain 3 0 - 6 0 mg. N a per 1 0 0 g.
can contents, or only about 5 - 1 0 % that of c o m m e r c i a l packs.
B . DlSCOLORATIONS AND APPEARANCE 1. Color
M u c h attention has b e e n given to the pigments in salmon and tuna and their changes in heat processing. Oxyhemoglobin is the chief pigment of raw t u n a meat. I n storage, in freezing, and on deaeration, m e t h e m o -globin is formed from oxyhemo-globin. I n h e a t processing, b o t h pigments are converted to hemochromes, the pigments that determine t h e color of the c a n n e d fish ( N a u g h t o n et al., 1 9 5 6 , 1957, 1 9 5 8 ; D u a n e B r o w n and T a p p e l , 1 9 5 7 ) .
Methods for evaluating the odor of raw and cooked salmon m e a t w e r e evolved b y S c h m i d t and Idler ( 1 9 5 8 ) with a view to predicting the color of the canned salmon from that of the raw flesh. No definite relation was found, however, b e t w e e n p H of the canned tuna liquid and freshness of the raw product ( K o c h i and Shitoku, 1 9 5 7 ) .
2. Browning
T h e muscle of m a n y white-fleshed fish shows a brown discoloration on heating. I n some species like lingcod, lemon sole, etc., the browning
is m u c h more pronounced than in others, e.g. halibut. T a r r showed ( 1 9 5 2 ) that this brown discoloration was due to c h e m i c a l interaction of the muscle protein with the reducing aldehyde group of sugars; this type of reaction is also observed in other foods and is known as the Maillard re
action. I n canning this reaction often causes concern, as it m a y not only induce an undesirable flavor and a p p e a r a n c e of the product, b u t also affect its nutritive value. Part of the protein m a y b e rendered insoluble and indigestible (Anonymous, 1 9 5 4 e ) .
T h e reaction r a t e of this browning is e n h a n c e d b y increased temper
atures; this is one of the main reasons for reducing h e a t processing to the lowest possible level for m a n y fish products. Spilde ( 1 9 5 1 ) has presented graphs showing the browning of fish paste as a function of temperature and time at various p H values.
T h e chemistry underlying the browning of fish m u s c l e in canning has been thoroughly investigated b y Tarr. H e determined the degree of browning b y measuring the reflectancy of fish flesh after heating to 1 2 0 ° C . ( 2 4 8 ° F . ) for 1 hour. Using a spectrophotometer, with an artificially pre
pared "white" surface as standard reference, h e related amount of sugars in the fish muscle ( T a r r , 1 9 5 3 ) to degree of browning. O f the sugars normally present in fish muscle—glucose, ribose, and deoxyribose—ribose was shown to b e the most active reagent in the browning reaction ( s e e T a b l e I I ) .
TABLE Ila
F R E E RIBOSE CONTENT AND REFLECTANCY OF HEATED FISH MUSCLE
Species
Free ribose (in % of wet muscle)
Reflectancy (in % of standard)
Halibut 0.035 63
Albacore tuna 0.007 57
Lemon sole 0.14 40
Lingcod 0.55 33
a Condensed from Tarr (1952).
Glucose is so m u c h less active than ribose that approximately 5 times as m u c h is n e e d e d to cause the same degree of browning. Deoxyribose, a constituent of the deoxyribonucleic acid of cell nulei, is present only in negligible quantities. H e n c e ribose was thought to b e the principal or even the exclusive agent in the browning reaction ( T a r r , 1 9 5 2 ) . Most of the ribose is believed to b e b o u n d in nucleotides or nucleic acids. I f halibut muscle is heated, there is little loss of total pentose ( a s deter
mined from a trichloroacetic acid e x t r a c t ) ; in lingcod muscle, however, 1 hr. of heating at 1 2 0 ° C . produced an appreciable loss of total pentose, but this loss does not increase on further heating. W h e n 1 mg. ribose was added to either halibut or lingcod muscle, none could b e recovered
after heating. H e n c e it is thought that free rather than bound ribose is the cause of browning. E v e n the ribosides guanosine, uridine, and adenosine did not increase browning in fish muscle as m u c h as ribose itself ( T a r r , 1 9 5 3 ) .
Ribose is furthermore freed from ribonucleic acid and similar com
pounds b y the enzyme riboside hydrolase ( T a r r , 1 9 5 4 a ) . T h e r e is little ac
tivity of this enzyme in meat, b u t it induces release of ribose in fish after death. Moreover, there are differences in the activity of this enzyme;
lingcod, rock cod, and flounder, but not halibut, can produce sufficient ribose ( 0 . 0 2 - 0 . 0 8 % during storage at 0 ° C . ) to allow browning on sub
sequent heating. Riboside hydrolase preparations from lingcod or rock cod added to halibut and salmon caused the same amount of ribose to b e formed during cold storage as in the quickly browning species ( T a r r and Bissett, 1 9 5 4 ) . In experiments with J a p a n e s e white fish meat, Ono and N a g a y a m a ( 1 9 5 8 - 1 9 5 9 ) found that browning was accelerated with de
creasing freshness of the meat, accompanied b y increasing contents of free reducing sugars; D-glucose was again found less effective than ribose in accelerating browning, b u t on the whole, a factor other than sugar was suggested as b e i n g most responsible for browning.
Finally, T a r r and Bissett ( 1 9 5 4 ) looked for means to remove the cause of the Maillard reaction. L i g h t cooking of the flesh and removal of the exuded liquid before canning will eliminate some of the sugar ( a n d flavor substances at the same t i m e ) , b u t some ribose will remain to react with proteins during heating. Another possibility is to b r e a k down the ribose to a derivative inactive in the Maillard reaction. In a search for a ribose oxidase in yeasts, molds, and bacteria, a preparation from Lactobacillus pentoaceticus was obtained which removed all ribose within 2 days at 0 ° C . when added to the fish. T r e a t e d fish showed no browning on heating.
It is not reported whether this procedure has gained practical application.
3. Iron Sulfide Discoloration
Less universal but more pronounced is the black or bluish discolora
tion due to the formation of F e S . This is a very old problem in canning.
Hess ( 1 9 5 6 ) cites a paper b y M c P h a i l and Briere ( 1 8 9 7 ) on the formation of iron sulfide in Canadian c a n n e d lobster as the first paper on this subject.
Through a long series of investigations it is now firmly established that F e S is formed b y a reaction of iron with free sulfides, particularly H2S in the fish. T h e main source of the iron is the insufficiently protected steel plate of the can, although in some cases the food itself m a y contain a high percentage of iron, or m a y have picked it up from the canning equipment or from the water. T h e latter source can usually b e easily
re-moved b y using aluminum or stainless steel e q u i p m e n t and a pure water source; w h e n the food itself contains a high amount of iron, as in the case of abalone, a wholly satisfactory procedure is not yet available.
T h e trend to reduce the tin coating of can materials ( s e e Section I V , C ) has created n e w problems, since b o t h electrolytic tin p l a t e cans and cans m a d e from previously l a c q u e r e d tin plate m a y offer insufficient pro
tection of the steel base. T h u s Mathiesen found ( 1 9 5 4 ) that electrolytic tin plate, unless extremely heavily tinned, is unsuitable for canning fish balls. A l a c q u e r coating may, in most cases, offer sufficient protection, e.g., by the extra strip of l a c q u e r along the inside of the seam. Important progress has b e e n achieved b y the introduction of the protective C-en-amel. Its protective action is b a s e d on the substitution of iron b y zinc, to induce the formation of white ZnS instead of the unsightly F e S .
W h i l e all these protective measures m a y b e effective in reducing the iron available for the reaction, it can in no w a y influence the presence of the other factor, the free sulfides. T h e s e are usually formed b y spoilage processes. In m a n y fish products, iron sulfide discoloration takes place in the can and m a y b e ascribed to preprocess spoilage. Certain rapidly spoiling seafoods, such as young eels, therefore require a combination of special precautions such as the use of only perfectly fresh r a w material, rapid handling, packing in C-enameled cans with paper p a r c h m e n t lining, and the application of acid ( B o u r y , 1 9 5 2 ) .
A detailed investigation into the causes of iron sulfide blackening was prompted recently b y its occasional o c c u r r e n c e in canned tuna.
T h o u g h seldom found in grated or flake-style tuna, it m a y suddenly ap
pear in solid packs in a whole b a t c h within 2 4 hours of processing ( H e s s , 1 9 5 6 ) . T h e discoloration occurs on the inside surface of the can and is concentrated in the area adjacent to the h e a d space. I t is harmless but unsightly and creates consumer prejudice. Another type of sulfide staining m a y form under the surface of the e n a m e l coating of the can, b u t is less objectionable as it does not affect the a p p e a r a n c e of the product.
Protein sulfur m a y have b e e n subject to c h e m i c a l alteration during cold storage previous to canning. I n c o m p l e t e b u t c h e r i n g m a y leave re
mains w h i c h favor sulfur production in t h e fish. I f precooking of the fish, which normally would drive volatile sulfide from the fish, is shortened, some H2S will b e retained in the fish to b e packed. D e l a y in the canning line will increase the risks of spoilage and m a y increase the amount of sulfide found. Processing time and temperature also have a b e a r i n g on sulfide formation ( P i g o t t and Stansby, 1 9 5 5 ) .
T h e iron of the can is accessible through tiny pores or cracks in the tin and the enamel coating. Metallic iron does not r e a c t with sulfides
under the conditions prevailing in a can. I t is possible that it is converted into a ferrous conpound or ion, which combines with sulfide to form black F e S .
T h e sulfide content of canned tuna was determined in a v a c u u m gauge, b y puncturing the cans when covered with 2 0 % zinc a c e t a t e solu
tion, so that all H2S in the h e a d space was immediately b o u n d ( P i g o t t and Stansby, 1 9 5 6 a ) . W h e n these determinations were compared for tuna b e fore and after precooking, it appeared that all free sulfide is driven from the m e a t in this process. Spoilage either before or after precooking did not increase discoloration more than that observed in freshly canned tuna, unless the cans were subsequently stored for extended periods.
On the other hand, albacore tuna held in cold storage for 14 months did show discoloration. This was more pronounced when the fish had been spoiled before freezing. L o c a l albacore, which was not frozen, showed no blackening; yellowfin from the Pacific showed some and al
b a c o r e imported from J a p a n serious dicoloration (Pigott, 1 9 5 6 ) .
As precooked tuna does not contain free sulfides, these must b e formed during retorting a n d / o r cooling. Sulfide increases with retorting and most of it is found as H2S in the head-space gas. T h e addition o f sulfide to the tuna did not increase blackening. Evidently the state of the iron is also a factor. I n the p r e s e n c e of sulfide, the discoloration is due to F e + + ions.
Discoloration m a y develop during the cooling phase and its intensity is enhanced b y increasing cooling temperature ( P i g o t t and Stansby, 1 9 5 6 b ) . Neither the amount nor the kind of salt or oil had a marked effect on this discoloration ( P i g o t t and Stansby, 1 9 5 7 a ) . So long as there is no free liquid, no blackening was found in p a c k from tuna, but it appeared promptly in packs which contained either salt and free oil, salt and free water, or free water alone. F r e e fatty acids in oil also caused discoloration, but only when present in abnormally large quantities. T M A and T M A O affect the amount of tin going into solution in canned herring ( J a k o b s e n and Matthiesen, 1 9 5 4 ) . This aggravates tin staining b u t not F e S forma
tion. O t h e r c o m m o n additives such as glycerin, tartaric acid, sucrose, glucose, and glutamate have no effect; nor do C u C l , C u C l2, and S n C l2
salts, b u t b o t h cupric and cuprous ions cause can corrosion which ulti
mately leads to pinhole leaks.
T h e s e investigations were climaxed b y a search for the factor w h i c h caused ferrous ions to b e formed ( P i g o t t and Stansby, 1 9 5 7 b ) . W h e n m e a t and oil from b l a c k e n e d tuna packs w e r e canned separately, the blacken
ing occurred in the m e a t pack, b u t not when the oil was added to non-b l a c k e n e d tuna. Composition of the head-space gas showed no clear rela
tion to incidence of discoloration, and p H had very little influence outside
normal amplitudes of variation. In c o m m e r c i a l packs b l a c k e n i n g oc
curred only in solid packs, b u t not in flake packs. T h e flake p a c k absorbs so m u c h oil that no free oil is left in the can. F l a k e packs with free oil showed the same incidence of blackening as the solid packs, w h i c h was confirmed b y earlier observations. B l a c k e n i n g was never found where meat or liquid was in permanent contact with the can during the cooling period.
H e n c e , in some varieties of tuna, there is a relatively heat-stable sub
stance contained in meat, or possibly developed and liberated during frozen storage, which can convert F e to the ferrous form which reacts with the sulfide present. This reaction, however, takes p l a c e only if a salt-containing liquid such as oil or brine is free and continuously in contact with the inner surface of the can during retorting. As the blackening gradually develops during cooling, it can b e prevented b y turning the cans over at a stage just before the F e S is deposited. T i n sulfide would then b e formed in the covered area. I t is possible to eliminate the dis
coloration in otherwise blackening packs b y inverting the cans 5 - 2 0 min.
after initiating the cooling. Unfortunately, this simple method is imprac
ticable in canneries where j u m b l e d loading of retorts is practiced.
4. Other Discoloration in Tuna
T h e normal pink color of canned tuna depends on one or two hemo-chromes, with bindings to globin a n d / o r to nicotinic acid. T h e pigment is heat-stable, b u t m a y b e oxidized to a tan color during cooking. T h e pink color can therefore b e preserved b y introducing a reducing agent before cooking, or b y exclusion of air ( D u a n e B r o w n and T a p p e l , 1 9 5 7 ) .
T h e pink color, however, can b e seriously overshadowed b y a brown pigmentation, when frozen tuna is used. This is caused b y the separation of hemoglobin which diffuses through the meat. This discoloration is markedly reduced, if the fish is frozen in water or brine instead of in air
(Anonymous, 1 9 5 2 i ) .
Finally, these same h e m e pigments m a y occasionally initiate serious green discolorations. This greening is similar to that which occurs in the m e a t of Pacific yellowfm tuna. J a p a n e s e workers found a correlation b e tween amount of myoglobin in the tuna m e a t and liability towards green
ing. T h e y were able to predict greening with 8 5 % accuracy from the hue of the fresh meat. Certain differences in moisture, vitamin B , and minerals were found b e t w e e n normal and green tuna ( H i r a o et al., 1 9 5 8 ) .
T h e chemistry underlying green discoloration is still largely obscure.
It appears to depend on the redox potential in the canned fish, w h i c h would explain its correlation with fat oxidation ( D u a n e B r o w n et al.,
1 9 5 8 ) . T h e r e seems to b e some connection with the denaturation of the pigment-bearing proteins (Naughton et al, 1957, 1 9 5 8 ) . Dollar et al
( 1 9 6 1 ) , however, found such correlations largely obscured b y individual variations, but were able to exclude microbial spoilage as a possible cause.
F o r further discussion see Chapter 4, this volume.
5. Struvite Crystals
T h e formation of magnesium ammonium phosphate hexahydrate ("struvite") crystals is frequently found in c a n n e d crustaceans ( c r a b , lobster, shrimps), but occasionally also in salmon, tuna, flaked cod and haddock, and roe. Although these crystals have no taste or odor and are otherwise quite harmless, they are often mistaken for glass fragments b y the consumer. This evokes aversion which has gone as far as threatening the producer with legal action ( M c F e e and Swaine, 1 9 5 3 ) .
F o r a discussion of struvite formation in shellfish, see Chapter 8A, this volume. Magnesium (usually from the sea water used in various opera
tions) combines with ammonia generated from the fish muscle protein during heat processing to form the struvite. This resultant compound gradually crystallizes. T h e p H is important, the phosphate being soluble in small amounts below p H 6.5 and in larger amounts below p H 4
( T a n i k a w a et al, 1 9 5 9 ; see also Kizevetter, 1 9 5 4 ) .
A large n u m b e r of substances have b e e n tested for ability to inhibit growth of these crystals ( L e Roux et al, 1 9 5 1 ) . Most efficient w e r e lecithin, emulsified oil, sodium sulfide, and sodium hexametaphosphate.
W h e n using this latter substance, T a n i k a w a et al. ( 1 9 5 7 ) observed a pale yellow discoloration, a bitter taste, and a whitish precipitate when more than 0 . 2 8 % of this compound was added to canned salmon. In model experiments they demonstrated that struvite crystals w e r e formed slowly in media of high viscosity (such as the juice formed in canned salmon or m a c k e r e l ) . O n c e formed, viscosity has little influence on their growth rate. M a n y small crystals are formed in solutions having a high concentra
tion of peptone, few but large crystals in solutions of gelatin, glycerin, agar, and edible oil.
I t was observed microscopically that the formation of the crystals begins during cooling after h e a t processing when the medium reaches a temperature of 8 0 - 7 0 ° C . ( 1 7 6 - 1 5 8 ° F . ) , the optimum for formation being 7 0 - 6 0 ° C . ( 1 5 8 - 1 4 0 ° F . ) . T h e freshness of the raw material affects the crystal formation, but less than does the cooling rate. T h e use of sodium hexametaphosphate as a sequestering agent is promising, as it is an almost neutral salt with little or no taste and has b e e n safely used in foods since it is rather insoluble. In the presence of N a C l it will hold 3 times as m u c h
struvite in solution as either N a C l solution or fish broth ( M c F e e and Swaine, 1 9 5 3 ) .
6. Curd
Curd is the n a m e for a protein coagulate often found on top of the fish in canned salmon and mackerel, species which are generally canned without precooking. A brine-soluble protein contained in the muscle exudes and coagulates during processing. As the water-binding properties of the muscle proteins are impaired b y freezing denaturation, curd forma
tion is especially serious wherever frozen salmon or mackerel are canned.
It has long b e e n known that brining the cut sections of defrosted salmon prior to canning materially lessens curd formation. T a n i k a w a et al. ( 1 9 5 2 ) found, b y measuring the amount of h e a t coagulable protein extracted from the muscle b y brines of different concentration, that treat
ment of cut m a c k e r e l with 1 0 - 1 5 % brine for 2 5 - 3 0 min., followed b y washing, was most effective. D a s s o w and Craven ( 1 9 5 5 ) studied the effect of brine dips and tartaric acid dips on thawed red salmon and Coho salmon steaks prior to canning. W i t h brine, a 10-min. dip in 7 0 % saturated brine was most effective and produced significant reduction in curd, b u t this treatment was less effective than a 1-min. dip in 5 % tartaric acid solution. Tartaric acid also prevented sticking of the m e a t to the can m u c h better than brine did. Seagran ( 1 9 5 6 ) studied the effect of salt and p H in further detail. H e measured retention of protein-containing fluids in salmon muscle b y dipping slices in 0 - 2 6 % brine solutions or b y sus
pending m i n c e d samples of thawed salmon m e a t in brines of various p H and centrifuging. T h e imbibing p o w e r of muscle proteins is influenced b y salt content and p H , and exhibits a zone of minimum effect correspond
ing to the isoelectric zone of the fish muscle proteins, i.e. p H 5.2-6.3. T h e retention of fluid in processing, with corresponding curd formation, de
pends on the liquid-binding power of the proteins at p H 6.5 and above, brought a b o u t b y the presence of 2 - 5 % salt in the fish. T h e amount of brine-soluble proteins in salmon muscle can of course vary widely be
tween species, and even b e t w e e n localities within one species, b u t m a y also c h a n g e considerably with the post-mortem changes. This was elucidated in a series of experimental packs b y S c h m i d t and Idler ( 1 9 5 5 ) . Despite w i d e individual variations, lengthy fresh storage time increased curd formation. This was m u c h more pronounced in sockeye than a blue-b a c k salmon. T h e increase in curd is thought to blue-b e related to the formation of soluble proteins during storage.
F o r frozen fish, it is necessary to minimize protein denaturation by short storage times and low storage temperatures. F o r r a w fresh market fish, it is necessary to minimize autolytic changes b y good preservation.
7. Honeycombing
H o n e y c o m b i n g is often encountered in canned tuna. I t consists of pits in the fish tissue, usually b e t w e e n the m e a t layers, and is seen after pre
cooking when the fish is taken apart on the cleaning table; sometimes the pits are also seen on cut surfaces. A cross section of extensively pitted tissue has an appearance suggesting an empty h o n e y c o m b . T h e phenom
enon usually starts in the n a p e region and proceeds tailward. Another starting point is the inside of the fish near the b a c k b o n e .
In canning practice, h o n e y c o m b i n g is always considered a sign of
In canning practice, h o n e y c o m b i n g is always considered a sign of