Tests LIONEL

Freshness Tests

LIONEL FÄRBER

Seafood Research Laboratory, The G e o r g e Williams Hooper Foundation, University of California Medical Center, S a n Francisco, California

I. Introduction 66 II. Organoleptic or Sensory Tests 66

III. Physical Methods 68 A. Texture Changes 68 B. Refractive Index 69 C. Electrical Conductivity 69

D. Optical Tests 69 E. Surface Tension, Viscosity, and Internal Friction 70

IV. Physicochemical Methods 71

A. pH 71 B. Buffering Capacity 72

C. Oxidation-Reduction Potential 73 D. Freezing Point Depression 73

V. Chemical Methods 73 A. Volatile Basic Nitrogen Compounds 73

B. Volatile Acids 78 C. Succinic Acid 80 D. Indole and Skatole 80 E. Hydrogen Sulfide 81 F. Carbonyl Compounds 81 G. Steam-Volatile Oxidizable Substances 82

H. Reducing Substances That Are Volatile at Room Temperature . . . 83

I. Histamine, Histamine-like Substances, and Others 85

J. Iodimetric Titration and Iodine Uptake 85

K. Protein Hydrolytic Products 86 L. Nucleotides and Derivatives 88 M. Fat Spoilage Methods 89 N. Miscellaneous Methods 92 VI. Biological Methods 94

A. Nitrate Reduction 94 B. Oxygen Consumption 94 C. Dye Reduction Tests 94 D. Bacterial Counts 96 E. Enzyme Activity Tests 97 VII. Concluding Remarks 97

References 99 Note Added in Proof 126

65

66 L I O N E L F A R B E R

I. Introduction

One of the primary aims of investigators of food spoilage has b e e n to find some m e t h o d b y w h i c h deteriorative changes could b e accurately and quantitatively measured throughout their course. An independent, objective scale of deterioration would m a k e available an impersonal reference standard that, when supplemented with sensory observations, would permit judgments of greater reliability than those b a s e d on sub­

jective evaluations alone. A brief survey of various appraisal methods for objective testing of the quality of fish, raw unfrozen, frozen, or otherwise processed, as well as an evaluation of their particular merits, was m a d e by Piskur ( 1 9 5 6 ) .

Several earlier reviews on the evaluation of freshness in seafoods as well as on the spoilage of fish are available, including those of Note- varp et al ( 1 9 4 2 ) , Hjorth-Hansen ( 1 9 4 3 ) , Allison ( 1 9 4 8 ) , R e a y and Shewan ( 1 9 4 9 ) , Partmann ( 1 9 5 1 ) , T a r r ( 1 9 5 4 , 1 9 5 5 a ) , T o m i y a s u and Zenitani ( 1 9 5 7 ) , W o j c i e c h and V a r e l a ( 1 9 5 8 ) , and Montefredine and T e s t a ( 1 9 6 0 ) .

In the present review are discussed as m a n y as p r a c t i c a b l e of the different types of methods that have b e e n proposed and investigated.

II. Organoleptic or Sensory Tests

T h e oldest and still most widespread means o f evalulating the accept­

ability and edibility of fish are the senses: smell and sight, supplemented b y taste and touch. T h e reasons for the preferential use of sensory tests are obvious: no special laboratory e q u i p m e n t is needed; the fish can b e examined w h e r e v e r they h a p p e n to b e ; the tests can b e carried out quickly; and m a n y samples can b e evaluated in a relatively short time.

T h e s e obvious advantages, however, are to a great extent counterbal­

anced b y a n u m b e r of disadvantages inherent in the organoleptic m e t h o d that significantly detract from its usefulness. T h e use of the senses is a subjective procedure and at b e s t only very roughly quantitative. T h e impressions registered are the result of the interaction of a n u m b e r of physiological, psychological, environmental, and even e c o n o m i c factors, including state of health, personal prejudices, preferences and interests, sensory acuity, freedom from disturbing and influencing conditions in the examining environment, and motives of possible profit or loss. T h e demand upon the senses b e c o m e s most critical and difficult w h e n they are required to distinguish and assess the so-called borderline stage of freshness or the stage of incipient spoilage ( t h a t is, when a fish sample to b e judged is in the last stages of freshness or in the first stages of

s p o i l a g e ) . T h i s task is well recognized as a difficult one, w h e r e even experienced judges often differ in their evaluations.

T h e criteria associated with freshness of fish have b e e n known for m a n y years. A m o n g the early studies is that of Anderson ( 1 9 0 8 ) , who accurately and carefully described the various criteria. H e recognized, among other signs, condition of the eyes, texture of the flesh, odor, and appearance of the abdominal walls. Stansby ( 1 9 4 4 ) and the Ohio D e p a r t m e n t of Health ( 1 9 5 3 ) have also discussed the problem of the organoleptic evaluation of fish from a n u m b e r of points of view. Boury ( 1 9 4 5 ) has discussed the sensory j u d g m e n t of canned fish. T h e impor­

tance of sensory judgment as a primary standard with which to compare other tests has b e e n pointed out b y Allison ( 1 9 4 8 ) .

T h e uncertainty and difficulties of sensory judgment of fish freshness have b e e n recognized b y those in the field. Castell et al. ( 1 9 5 6 ) have expressed similar ideas b y stating that "in actual p r a c t i c e the grading of fresh fish is, and for some time will remain, an art and not a science."

T o overcome some of the disadvantages associated with sensory judg­

ment, attempts h a v e b e e n m a d e to minimize the personal uncertainties and to overcome the qualitative nature of the test b y means of trained or experienced panels to judge the samples, and b y the elaboration of numerical systems of scoring and recording the sensory judgments.

Young ( 1 9 3 8 ) has described the use of a taste panel for the evaluation of halibut samples. Griffiths and S t a n s b y ( 1 9 3 4 ) , Castell et al. ( 1 9 5 6 , 1 9 5 8 ) , Hansen ( 1 9 5 6 ) , Shewan et al. ( 1 9 5 3 ) , and Shewan and E h r e n b e r g ( 1 9 5 6 ) have described systems of grading and of numerical scoring of r a w and cooked fish b a s e d on the evaluation of such factors as odor, general appearance, taste, and texture of raw and cooked fish. Score sheets and a point system for the numerical scoring of the samples have b e e n elaborated as a basis for the sensory judgments. Soudan et al.

( 1 9 5 7 ) h a v e also described a special system for the organoleptic estima­

tion of fish spoilage. T h e relation b e t w e e n sensory and c h e m i c a l tests has b e e n discussed b y J e n s e n ( 1 9 5 6 ) ; some general aspects of sensory testing procedures h a v e b e e n discussed b y Harries ( 1 9 5 3 ) .

All of these systems are b a s e d on the summation and numerical representation of a n u m b e r of sensory judgments of odor, texture, general appearance, etc., of r a w fish, and of the taste o f cooked fish. T h e y have proven fairly serviceable, providing reasonably good sensory yard­

sticks o f quality and freshness of fish samples within the limitations of such procedures.

A variation of the sensory test, in w h i c h the changes in t h e appear­

ance of the lenses of fish eyes are used as the criterion of the condition

6 8 L I O N E L F A R B E R

of the fish, has b e e n described b y L o v e ( 1 9 5 4 , 1 9 5 6 ) . T h i s test has proven of some value under controlled conditions, b u t its general useful­

ness and applicability are influenced and limited b y the treatment of the fish, such as freezing and other storage conditions.

W h e r e a s sensory or organoleptic tests of the state of freshness of fish and fish products have proven of general value and are used widely throughout the world (particularly for samples of either unquestioned freshness or definite s p o i l a g e ) , their value is markedly diminished and they b e c o m e less exact and less consistent when applied to fish at the stage b e t w e e n the end of freshness and the beginning of spoilage, t h e area of so-called incipient spoilage. I t is precisely in this state of pres­

ervation that the sensory judgments of individuals b e c o m e m o r e vari­

able, m o r e dependent on subjective factors, and therefore less reliable.

E v e n experienced persons have great difficulty in making an accurate and quantitative estimation of the condition of fish in this state. T h e line dividing fish that are still fresh from those with some early signs of spoilage is not very well defined and is most often subject to differ­

ences in personal opinions. As a result, investigators for about three- quarters of a century have attempted to devise tests that would b e less dependent on personal factors and subjective opinion. This review is essentially an account of the attempts through the years to develop m o r e objective criteria, b a s e d on physical or c h e m i c a l tests, that would overcome those disadvantages inherent in t h e sensory, m o r e subjective criteria.

III. Physical Methods

Several studies have b e e n directed toward the development of prac­

tical deterioration tests b a s e d on changes in the physical properties of the fish flesh. I n general these tests have not proven applicable, although a few, under certain limited and circumscribed conditions, have shown some correlation with the freshness of fish.

A. T E X T U R E C H A N G E S

F o r b e s ( 1 9 2 6 - 1 9 2 7 ) reported studies on the effect of various treat­

ments on the tensile strength of fish muscle. H e found a steady decrease in the breaking point of muscle strips after rigor, the rate of decrease varying with the temperature.

In 1 9 3 1 T a u t i et al. described a m e t h o d for measuring the force n e e d e d to deform t h e surface of r a w fish. T h e degree of deformation and the time n e e d e d for the return of the surface after removing the force were used as a measure of the freshness of the fish. Y a m a m u r a

( 1 9 3 2 , 1 9 3 3 a ) reported further tests with this m e t h o d and on the cor-

relation of t h e loss of rigidity of t h e flesh with the increase in total volatile nitrogen content during spoilage. C h a r n l e y ( 1 9 3 6 ) and Charnley and Bolton ( 1 9 3 8 ) tested the firmness or softness o f c a n n e d salmon as a measure of its condition. T a k a h a s h i et al. ( 1 9 5 2 ) described experi­

ments on the relationship b e t w e e n the depth of penetration of a plunger into fish jellies and t h e changes in total volatile nitrogen and p H during spoilage. No marked c h a n g e in rigidity was found at the point w h e r e a m a r k e d a m m o n i a c a l odor was noted. L o v e ( 1 9 6 0 ) described a n e w pro­

cedure for optically measuring the texture o f fish m u s c l e and its changes during storage. Buttkus ( 1 9 6 3 ) recently described an apparatus for m e a s ­ uring texture changes b a s e d on the energy n e e d e d to cut muscle fibers.

B . R E F R A C T I V E I N D E X

T h e possible use of the refractive index to follow the changes in fish during storage and spoilage has b e e n studied, without m u c h success, b y Tillmans et al. ( 1 9 2 7 ) , B o u r y and Schvinte ( 1 9 3 5 ) , Riddell et al.

( 1 9 3 7 ) , and Sidaway ( 1 9 4 1 ) . A r e c e n t modification of this test has b e e n reported b y Proctor et al. ( 1 9 5 9 ) . T h e s e authors studied changes in t h e refractive index of the fluid of fish eyes during storage of w h o l e h a d d o c k under different conditions. T h e y reported a good correlation b e t w e e n the value of the refractive index and the organoleptic judgment. F u r t h e r tests of this procedure will determine its practical use for r a w whole or eviscerated fish.

C . E L E C T R I C A L C O N D U C T I V I T Y

T h e possible value of the determination of the electrical conductivity of fish flesh as a m e a s u r e of its condition has b e e n studied at various times. Tillmans et al ( 1 9 2 7 ) , Riddell et al. ( 1 9 3 7 ) , and L a b a r r e and F o u g e r e ( 1 9 4 2 ) reported that conductivity was of no practical value as an index of fish spoilage since it varied with storage time and tem­

perature, passing through a m i n i m u m during the early stages and rising again as definite spoilage developed. Y a m a d a and Kitano ( 1 9 4 8 ) , Ito et al. ( 1 9 4 9 ) , and Asakawa ( 1 9 5 6 ) reported that the electrical resistance of carp m u s c l e c h a n g e d stepwise during storage as the fish passed through the stages of pre-rigor, rigor, and post-rigor to definite spoilage.

Amano ( 1 9 5 4 ) concluded, however, that the p r a c t i c a l application of the test was unwarranted. T h e consensus o f the value o f this test is that it is not reliable for determination of the c h a n g e from fresh to not-fresh.

D . O P T I C A L T E S T S

A variety o f tests b a s e d on optical properties o f fish flesh and its products h a v e b e e n studied over the years. I n some instances favorable

7 0 L I O N E L F A R B E R

results have b e e n reported. T h e general conclusion, however, has b e e n that these tests depend upon too m a n y uncontrollable factors that reduce their value as reliable methods for measuring freshness. In 1933 L a n g ­ stroth found no direct relation b e t w e e n the absorption of white light and the amount of bad-smelling products in fish. B o u r y and Schvinte ( 1 9 3 2 , 1 9 3 5 ) , Hinard ( 1 9 3 2 ) , van de V e l d e ( 1 9 3 7 ) , Leskov ( 1 9 4 6 ) , Adamova and S p e c t o r ( 1 9 4 7 ) , and Wittfogel ( 1 9 5 2 ) studied the ultra-violet fluorescence and luminescence of dispersions of fresh and spoiled fish flesh. T h e y reported that changes in the color of the fluorescence oc­

curred during spoilage, b u t that such factors as the state of dispersion and the time of extraction affected the colors observed. L i m b u r g ( 1 9 5 1 ) studied the possibilities of determining the green or b l u e fluorescence appearing in fresh-water fish and caused b y bacterial activities. Quite fresh fish showed no fluorescence at all. N o conclusive results were obtained. T o m i y a m a et al. ( 1 9 5 5 ) reported a turbidimetric method for determination of the freshness of the cooked fish paste, "kamaboko." As the surface of the fish loaf underwent b a c t e r i a l decomposition, it b e c a m e more water-dispersible and h e n c e increased the turbidity of the surface washings.

Recently Kurtzman and Snyder ( 1 9 6 0 ) reported an optical test for estimating the freshness of crabmeat, b a s e d on the turbidity of alcohol extracts that developed upon the addition of saturated aqueous picric acid solution. This test is more suited for establishing the presence of spoilage rather than the degree of freshness.

E . S U R F A C E T E N S I O N , V I S C O S I T Y , AND I N T E R N A L F R I C T I O N

A n u m b e r of investigators h a v e attempted to follow the changes occurring during deteriorative storage of fish b y measuring several other physical properties. F o r example, Tillmans et al. ( 1 9 2 7 ) showed that surface tension did not change markedly during deteriorative storage.

T h e viscosity was of no value as an indicator of spoilage in fish flesh;

it underwent a cyclic change, increasing somewhat at first, dropping to a minimum, and finally, with the onset of definite spoilage, mounting again ( L a b a r r e and F o u g e r e , 1 9 4 2 ) . Hotani ( 1 9 5 1 ) studied the internal friction o f fish flesh b y a damping oscillation method. H e found that this quantity varied with the freshness of the samples, and that the test could b e used for frozen fish without defrosting. No further reports have apparently b e e n published on this method, and its significance for the detection of the early stages of spoilage remains undetermined. F r o m its similarity to the physical properties just mentioned, particularly viscosity, it is, however, somewhat doubtful if this m e t h o d could serve as a gen­

erally reliable index of incipient spoilage.

IV. Physicochemical Methods

M a n y reports have appeared during the past 3 5 years or so on the correlation b e t w e e n certain p h y s i c o c h e m i c a l measurements and the early stages of spoilage. T h e bulk of these has b e e n c o n c e r n e d with p H changes, and several have dealt with buffering capacity and redox potential.

A. P H

T h e r e has b e e n an extensive series of studies on the value of p H as a measure of fish spoilage. Although some of these reports h a v e b e e n contradictory or conflicting, the majority of them have b e e n consistent in their conclusion that p H has little or no significance as a reliable index of the state of freshness of a sample ( Y a m a m u r a , 1933b, and o t h e r s ) .

B e n s o n ( 1 9 2 8 ) , Poluektov ( 1 9 3 3 ) , L ü c k e and Geidel ( 1 9 3 5 ) , Okolov and Shavskii ( 1 9 3 6 ) , Zwilling ( 1 9 3 6 ) , Riddell et al ( 1 9 3 7 ) , Zakhar'evskii ( 1 9 3 9 b ) , B r a d l e y a n d B a i l e y ( 1 9 4 0 ) , Hjorth-Hansen ( 1 9 4 3 ) , E l l i o t t ( 1 9 4 7 ) , Sigurdsson ( 1 9 4 7 ) , Kondrup ( 1 9 4 8 ) , Rasmussen ( 1 9 5 0 ) , Cutting ( 1 9 5 3 ) , Higasa ( 1 9 5 3 ) , U c h i y a m a and Y o k o y a m a ( 1 9 5 3 a ) , Y a m a m o t o and Sonehara ( 1 9 5 3 ) , Luijpen ( 1 9 5 4 a , b ) , Simidu and Hibiki ( 1 9 5 4 b , c , d ) , V a r e l a and W o j c i e c h ( 1 9 5 6 ) , W o j c i e c h and V a r e l a ( 1 9 5 8 ) , and Zakhar'­

evskii ( 1 9 3 9 a , b ) h a v e reported that the p H varied initially and that it showed no real or significant correlation with the onset of spoilage in fresh fish (e.g., cod, haddock, perch, herring, halibut, and h a k e ) , salted fish

(e.g., herring, carp, pike, and r o c k ) , fish loafs, or shellfish. O n the other hand, van Deurs and Hoff-j0rgensen ( 1 9 3 6 ) proposed an upper limit of p H 7.5 for acceptability of cod fillets. Strohecker et al ( 1 9 3 7 ) reported that the p H increased in fish parallel to the increase in the content of volatile oxidizable substances of steam distillates. Charnley and G o a r d

( 1 9 4 2 ) suggested the use of p H as a measure of the freshness of fish mus­

cle, such as salmon. Finally, W o o d et al ( 1 9 4 2 ) and D y e r et al ( 1 9 4 4 ) maintained that the p H of the surface could b e used as a rapid and simple test for the degree of freshness of white-meat fish, such as cod and haddock. I n an extensive study o f the fishes o f the Adriatic Sea, Misericordia ( 1 9 5 4 ) established the upper limit for p H in spoilage of e a c h individual species and found that spoilage ran parallel to the mounting p H of the muscle tissue. Montefredine ( 1 9 5 5 ) also concluded that p H was a reasonably reliable indicator of the degree of freshness.

T h e consensus from the above studies is that the p H determination cannot b e used as a reliable index of the state of freshness or of the onset of spoilage, b u t that under certain restricted conditions p H limits for definite levels of spoilage m a y sometimes b e set. T h e usefulness of p H determination is often greatly restricted or vitiated b y its variability

7 2 L I O N E L F A R B E R

from sample to sample and b y its cyclic fluctuations during the storage process.

B . B U F F E R I N G C A P A C I T Y

I n 1 9 3 3 Stansby and L e m o n published a paper in w h i c h a differential titration to two levels of p H ( A and Β v a l u e s ) was suggested as a measure of the spoilage of haddock. T h e y ascribed the changes in values to the breakdown of proteins and the formation of protein decomposition products, mainly of a b a s i c nature. Griffiths and Stansby ( 1 9 3 4 ) elab­

orated on this m e t h o d and c o m b i n e d it into a single value. This proce­

dure was tested b y Riddell et al ( 1 9 3 7 ) , F i t z g e r a l d and C o n w a y ( 1 9 3 7 ) , and Cutting ( 1 9 3 8 , 1 9 5 3 ) . T h e s e authors reported that the changes in titration values for the two p H levels w e r e too irregular and variable to form the basis of a reliable spoilage test and that no correlation b e t w e e n organoleptic j u d g m e n t or bacterial content and the buffering capacity was found.

A decreased buffering capacity was noticed during spoilage of c o d muscle press juice; it was ascribed to a reduction of trimethylamine oxide

( T M A O ) to the free b a s e (Collins et al, 1 9 4 1 ) .

H a c k e r ( 1 9 5 0 ) has c h e c k e d the Stansby and L e m o n m e t h o d of deter­

mining buffering capacity of h a d d o c k muscle on the muscle of cod as well as of three fresh-water fish (trout, roach, and sunfish). T h e Β values w e r e found to increase for all four fishes, confirming t h e earlier findings, whereas t h e A values, in contrast to the previous observations, declined in all cases. I n separate tests it was found that lactic acid and T M A O h a d strong buffering capacity in the p H range 6 . 0 - 4 . 3 , b u t that trimethylamine ( T M A ) h a d none.

T h e m e t h o d was also tested on fresh-water fish b y B o s e and D u t t ( 1 9 5 4 ) . T h e y found that t h e titration values A and Β showed a rough correlation with the sensory grades b u t that they overlapped in other respects. T h e trend of the results for fish stored at higher temperatures was different from those found for fish stored in ice. As a result no definite limiting values could b e established for fish judged to b e spoiled.

T h e premise and application of this m e t h o d h a v e also b e e n discussed b y Nickerson and Proctor ( 1 9 3 5 ) and B e a t t y and G i b b o n s ( 1 9 3 7 ) .

T h e buffering capacity o f fish muscle and m u s c l e extractives was restudied b y S u y a m a and Tokuhiro ( 1 9 5 8 ) in order to test its value as a criterion of the d e g r e e of freshness of fish muscle. T h e m e t h o d proved less reliable, although titration to t h e B-level seems to have possibilities.

T h e muscles of elasmobranchs show a higher degree of buffering than those of the teleosts b e c a u s e of the a b u n d a n c e of T M A O .

C. O X I D A T I O N - R E D U C T I O N P O T E N T I A L

T h e possible correlation b e t w e e n changes in t h e redox potential and the sensorily determined condition of fish has b e e n studied a n u m b e r of times ( T i l l m a n s et al, 1 9 2 7 ; Riddell et al, 1 9 3 7 ; Zakhar'evskii 1 9 3 9 a , b ) . I t was established that t h e redox potential was of n o practical use as a spoilage indicator for fish. Proctor et al. ( 1 9 5 7 ) also found t h e redox po­

tential too variable to b e of use as a fish spoilage test.

D . F R E E Z I N G P O I N T D E P R E S S I O N

Tillmans et al ( 1 9 2 7 ) studied the depression of the freezing point of fish flesh during spoilage. T h e y found that the entire range of values was quite small and that the freezing point decreased to a minimum value at the b e g i n n i n g of spoilage and then increased on further spoil­

age, r e a c h i n g the original value of t h e fresh samples. T h e y concluded that its estimation was an unsuitable test of spoilage.

V. Chemical Methods

A. V O L A T I L E B A S I C N I T R O G E N C O M P O U N D S

1. Total Amount

M o r e data have p r o b a b l y b e e n a c c u m u l a t e d on this test than on any of t h e others suggested as spoilage indicators. F u r t h e r m o r e , there p r o b ­ ably are m o r e conflicting results for this test than for any others. F o r example, of 6 4 papers reviewed, 3 8 r e p o r t e d favorable results for the use of the total volatile bases as a spoilage indicator, 17 stated that it was of no use, and 9 presented data showing the variable nature of the test, w h i c h is useful for some species or products and useless for others.

T h e earliest report found on the use of a m m o n i a or volatile bases as an index o f spoilage was that of E b e r ( 1 8 9 1 ) , w h o suggested that the fuming of a dilute H C l solution in ether-alcohol could b e employed as a qualitative test for the spoilage of meat. Clark and A l m y ( 1 9 1 7 a , b , 1 9 2 0 ) studied a n u m b e r of possible methods and reported that the total volatile b a s i c nitrogen content increased during storage of shucked oysters and of white-meat fish. W e b e r and W i l s o n ( 1 9 1 9 , 1 9 2 0 ) studied the volatile b a s i c nitrogen compounds in c a n n e d sardines during decom­

position of small sea herring. H i n a r d ( 1 9 2 2 ) suggested the ratio of the content of total volatile b a s i c nitrogen to that o f total nitrogen as a useful index of fish spoilage. Tillmans and Otto ( 1 9 2 4 ) found that the total volatile bases increased with the onset of spoilage of such fish as cod, haddock, eel, a n d sea pike; they suggested an u p p e r limit of 3 0 mg. nitrogen p e r 1 0 0 g. for acceptability. Glassmann a n d

74 L I O N E L F A R B E R

R o c h w a r g e r ( 1 9 2 9 ) also reported favorable results for the volatile basic nitrogen content as determined with Nessler's reagent and suggested 2 0 mg. Ν per 1 0 0 g. as the upper limit for good fish. R e e d et al. ( 1 9 2 9 ) and Gibbons and R e e d ( 1 9 3 0 ) compared the production of volatile basic nitrogen from haddock, clams, and lobsters during autolysis with that during bacterial decomposition. T h e y found that only during the latter process were appreciable amounts of volatile nitrogen produced.

Y a m a m u r a ( 1 9 3 3 b ) and T a n i k a w a ( 1 9 3 5 ) also found the content of volatile basic nitrogen useful as a measure of spoilage, and suggested 3 0 mg. Ν per 1 0 0 g. as the upper limit for acceptability. F a v o r a b l e results for the increase of the total volatile basic nitrogen content with fish spoilage w e r e also found b y B o u r y and Schvinte ( 1 9 3 2 , 1 9 3 5 ) , B o u r y ( 1 9 3 4 , 1936, 1 9 3 7 ) , Dobrovskii and Novikova ( 1 9 3 5 ) , Salmon and L e G a l l ( 1 9 3 6 ) , Smorodintzev and Kruilova ( 1 9 3 6 ) , Zwilling ( 1 9 3 6 ) , Firsov ( 1 9 3 7 ) , L a b r i e and Gibbons ( 1 9 3 7 ) , Crooks and R i t c h i e ( 1 9 3 8 ) , Shewan ( 1 9 3 8 , 1939a, 1 9 4 2 ) , Holmov ( 1 9 3 9 a ) , Stansby et al. ( 1 9 4 4 ) , O t a ( 1 9 5 8 a , b ) , R i e m a n n ( 1 9 5 2 ) , T o m i y a m a et al. ( 1 9 5 2 ) , Pierangeli et al.

( 1 9 5 4 ) , T a n i k a w a et al. ( 1 9 5 5 ) .

L ü c k e and Geidel ( 1 9 3 5 ) suggested an upper limit for the beginning of spoilage of 3 0 mg. volatile basic Ν per 100 g. K i m u r a and Kiamakura ( 1 9 3 4 ) for salmon r e c o m m e n d e d volatile b a s i c nitrogen levels per 100 g.

of 10 mg. or less for fresh, 2 0 - 3 0 mg. for beginning of spoilage, and over 3 0 mg. for spoiled fish. T a n i k a w a and Akiba ( 1 9 5 5 ) for c r a b m e a t sug­

gested 2 0 mg. volatile basic nitrogen per 100 g. as the upper limit for fresh meat.

T h e content of total volatile bases has in general b e e n found a rather unsatisfactory indicator of spoilage. V a r i a b l e results have b e e n reported b y several authors, including B ö k m a n ( 1 9 1 8 ) , Tillmans et al. ( 1 9 2 7 ) , and Okolov ( 1 9 3 2 ) for herring; Poluektov ( 1 9 3 3 ) and Pershin ( 1 9 3 5 ) for salted herring; B e a t t y and Gibbons ( 1 9 3 7 ) for white-meat fish; T a n i ­ k a w a ( 1 9 3 8 a , b ) for carp, salmon, sardines, crabs, and oysters; L a b a r r e and F o u g e r e ( 1 9 4 2 ) for salted and dried cod; T a r r ( 1 9 4 2 ) for canned herring; Boury ( 1 9 4 5 ) for canned fish; Rasmussen ( 1 9 5 0 ) for m i n c e d fish; T o m i y a m a and Y o n e ( 1 9 5 3 ) for "kamaboko" (fish l o a f ) ; U c h i y a m a and Yokoyama ( 1 9 5 3 ) for fish cakes; F ä r b e r ( 1 9 5 2 ) and F ä r b e r and F e r r o ( 1 9 5 6 ) for dark-fleshed, raw, and canned fish; Vaisey ( 1 9 5 6 ) for cod; and Moorjani et al. ( 1 9 5 8 ) for fresh-water fish.

Clark and Almy ( 1 9 1 7 a ) , although they observed some increase in the content of volatile bases during fish spoilage, nevertheless suggested that the determination of the content of other volatile substances asso­

ciated with the development of off-odors offered a more promising means

to assess spoilage. T h e conflicting and contradictory results that have b e e n reported, partly with m e a t other than that of fish, are presumably explained b y differences in composition, bacterial flora, handling methods, etc., b e t w e e n commodities ( s e e further V o l u m e I, C h a p t e r 1 4 ) . W a k s - man and L o m a n i t z ( 1 9 2 5 ) found that if little or no carbohydrate were present, t h e content of ammonia or volatile bases increased appreciably during m e a t spoilage, but that if carbohydrates were present, protein decomposition with ammonia formation was depressed. A similar phe­

nomenon was reported b y S'midu and Hibiki ( 1 9 5 7 ) .

K a w a b a t a ( 1 9 5 3 b ) found that during the spoilage of such red-meat fish as albacore tuna and mackerel no marked increase in total volatile bases occurred, in contrast to the m a r k e d increase during spoilage of white-meat flatfish. ( F o r the latter he suggested 3 0 mg. of volatile basic nitrogen per 1 0 0 g. as the upper limit for freshness.) B o s e and D u t t

( 1 9 5 4 ) reported that for spoilage of fresh-water fish at temperatures of 8 0 - 9 0 ° F the content of volatile bases showed a rough correlation with the freshness, b u t that at 3 0 - 4 0 ° F no correlation was found b e t w e e n the sensory j u d g m e n t and the content of volatile bases. F ä r b e r and Ceder- quist ( 1 9 5 3 ) found that the content of volatile b a s i c nitrogen was of some value for white-meat fish spoilage b u t not for that of red-meat fish. Buff a and Ambanelli ( 1 9 5 4 ) reported that total volatile basic nitro­

gen content could b e used to j u d g e the condition of raw unfrozen and frozen tuna and mackerel, but that it was not suitable as an indicator of the condition of the canned fish. This index of spoilage furnished a rea­

sonably accurate and rapid m e t h o d for appraising the keeping quality of cured fish products according to Indian studies (Pillai and Nayar, 1 9 5 7 ) , and it was found to b e superior to the m e t h o d b a s e d on determination of T M A b y V a l e n k a r ( 1 9 5 2 ) .

Wierzhchowski ( 1 9 5 6 ) found that the content of total volatile b a s e was useful for estimating the freshness of lean fish, such as cod, and suggested 3 0 - 4 0 mg. Ν per 1 0 0 g. as the upper limit for fresh-water fish and 6 0 mg. Ν per 1 0 0 g. as the limit for marine fish. However, the total volatile b a s e content was of no value for herring. L y t h g o e ( 1 9 1 3 , 1 9 3 8 ) and Fellers et al. ( 1 9 5 7 ) r e c o m m e n d e d t h e ratio of volatile basic nitrogen to the total nitrogen as a useful index of the quality of fish.

Sato ( 1 9 5 8 , 1 9 6 0 ) studied the volatile b a s i c nitrogen and amino nitrogen in fresh market fish and the changes in spoilage. E m p l o y i n g Conway's microdiffusion methods, h e found this m e t h o d useful b u t noted a consistent difference b e t w e e n b o t t o m fish and surface fish, the latter always having a lower value for volatile amino nitrogen in relation to volatile b a s i c nitrogen.

7 6 L I O N E L F A R B E R

2. Ammonia

I n the paper b y L y t h g o e ( 1 9 1 3 ) referred to earlier, the ratio b e t w e e n ammonia and total nitrogen was found to b e an adequate test of decom­

position in fish. Additional data supporting this conclusion w e r e presented b y L y t h g o e in 1 9 3 8 . S i n c e then very little further information has ac­

cumulated along this line.

Only in 1 9 5 2 did O t a and Nakamura establish that the quantity of ammonia in r a w and in precooked m e a t of various fishes h e a t e d under pressure was proportional to the freshness of the raw meats. T h e y inferred that the grade of freshness of the original raw material could b e estimated on the basis of the ammonia content found in canned fish products.

No real difference was established b e t w e e n thawed frozen and raw unfrozen fish muscle as to increase in ammonia, whereas histamine forma­

tion was largely suppressed in thawed muscle ( O t a and Kaneko, 1 9 5 8 ) . Differences prevail as to species and freshness stages in several Italian fishes ( T e s t a , 1 9 5 5 ) . No conclusive results were r e a c h e d as to the useful­

ness of this m e t h o d for determining degree of spoilage.

3. Trimethyfomine (TMA)

T h e situation for determining spoilage b y T M A is somewhat analogous to that of the other volatile b a s i c nitrogen compounds. A n u m b e r of re­

ports have appeared r e c o m m e n d i n g this method, others h a v e claimed it is of little value, and s o m e h a v e reported equivocal results. As with the volatile b a s i c nitrogen studies, the conflicting situation for trimethylamine is partly the result of the use of different species of fish with varying compositions and of the employment of different storage temperatures and conditions.

Poller and L i n n e w e h ( 1 9 2 6 ) reported that during fish spoilage tri­

methylamine oxide ( T M A O ) was reduced b y bacteria to trimethylamine.

Boury was one of the early investigators to report that trimethylamine could serve as an index of spoilage for r a w fish ( B o u r y and Schvinte, 1932, 1 9 3 5 ; Boury, 1 9 3 6 ) . T h e initial impetus to the use of trimethyl­

amine as a measure of spoilage, which has survived to the present, was provided, however, b y the first report of B e a t t y and Gibbons ( 1 9 3 7 ) , who also presented the rather simple C o n w a y microdiffusion technique for its determination. In later papers B e a t t y ( 1 9 3 8 , 1 9 3 9 ) and W a t s o n ( 1 9 3 9 ) showed the probable origin of T M A in cod muscle press juice

and worked out likely c h e m i c a l reactions for its formation from t h e precursor, T M A O . D y e r ( 1 9 4 3 , 1 9 4 5 ) described a colorimetric procedure for the determination of T M A ; this has b e e n modified b y Hashimoto and Akaichi ( 1 9 5 7 ) . T h e increase in T M A during spoilage has also b e e n studied b y H o l m o v ( 1 9 3 9 a , b ) , D y e r and M o u n s e y ( 1 9 4 5 ) , D y e r et al

( 1 9 4 6 ) , Proctor et al ( 1 9 5 0 ) , T a k a h a s h i et al ( 1 9 5 2 ) , Simidu and Hibiki ( 1 9 5 4 a , 1 9 5 5 ) , Castell and Triggs ( 1 9 5 5 ) , B a i l e y et al ( 1 9 5 6 ) , Horie and Sekine ( 1 9 5 6 ) , and S h e w a n and J o n e s ( 1 9 5 7 ) . Hess ( 1 9 4 1 ) sug­

gested that t h e value of the T M A determination could b e e n h a n c e d b y determining its increase in fish fillets after an incubation period. A measure of the keeping quality during subsequent storage could also b e obtained in this manner.

T M A is generally not found in fresh-water fish ( s e e V o l u m e I, C h a p t e r 6 ) ; consequently this m e t h o d cannot b e employed with this category o f fish.

A series of papers on the grading of whole and filleted fish b y the content of T M A has b e e n published b y Castell et al ( 1 9 5 8 ) , Hoogland ( 1 9 5 8 ) , and Castell and G r e e n o u g h ( 1 9 5 8 ) . F u r t h e r m o r e , a n u m b e r of reports have appeared in w h i c h it was found that t h e content of T M A was not a sensitive, reliable, or reproducible index o f fish spoilage.

E i t h e r the increase in T M A occurred during t h e latter stages of spoilage, or there was a variation in levels b e t w e e n species, or the ranges of values were so w i d e and overlapping as to preclude setting up definite stand­

ards, or no appreciable increase took p l a c e . A m o n g the latter studies the following m a y b e consulted: Tillmans and Otto ( 1 9 2 4 ) ; Tillmans et al ( 1 9 2 7 ) ; Okolov ( 1 9 3 2 ) ; B o u r y ( 1 9 3 4 , 1 9 4 5 ) ; Riddell et al. ( 1 9 3 7 ) ; Sigurdsson ( 1 9 4 7 ) ; Anderson and F e l l e r s ( 1 9 4 9 ) ; T a r r and N e y ( 1 9 4 9 ) ; F a r b e r ( 1 9 5 2 ) ; V e l a n k a r ( 1 9 5 2 ) ; K a w a b a t a ( 1 9 5 3 a ) ; Simidu and Hibiki ( 1 9 5 4 b , d ) ; E h r e n b e r g and S h e w a n ( 1 9 5 5 ) ; G o o d and Stern ( 1 9 5 5 ) ; Luijpen ( 1 9 5 4 a , b ) ; V a r e l a and W o j c i e c h ( 1 9 5 6 ) ; F ä r b e r and F e r r o

( 1 9 5 6 ) ; S h e w a n and E h r e n b e r g ( 1 9 5 7 ) ; L u i j p e n ( 1 9 5 8 ) .

I t has also b e e n suggested that T M A is a product of the early stages of spoilage (Collins, 1 9 3 8 ; Hess, 1 9 4 1 ) and that it m a y b e lost indis­

criminately during storage. Dussault ( 1 9 5 7 ) found that spoilage levels of T M A w e r e not the same for rosefish as for c o d fillets and that the T M A values w e r e a good basis for grading rosefish fillets, using differ­

ent spoilage levels from those that w o u l d b e applicable for c o d fillets.

T h e relation of b a c t e r i a l population to T M A content has b e e n studied b y T a r r ( 1 9 3 8 , 1 9 3 9 , 1 9 4 0 ) , S h e w a n ( 1 9 3 9 b ) , W o o d and B a i r d ( 1 9 4 3 ) , Neilands ( 1 9 4 5 ) , D y e r and W o o d ( 1 9 4 7 ) , and Robinson et al. ( 1 9 5 2 ) .

78 L I O N E L F A R B E R

Ronold and J a k o b s e n ( 1 9 4 7 ) found that T M A could also b e produced in canned fish b y a chemical reduction of the T M A O . Castell ( 1 9 4 9 ) and Castell and Greenough ( 1 9 5 7 b ) showed that nitrite suppressed T M A formation, whereas the tetracycline antibiotics did not.

4. Other Amines

Most of the determinations on volatile nitrogenous bases have either measured the total content of these substances or have involved the estimation of T M A with whatever amount of dimethylamine ( D M A ) may have been present. A few investigators have measured the amount of the individual amines formed during spoilage. Guareschi ( 1 9 1 7 ) and Tillmans et al. ( 1 9 2 7 ) found a positive test for carbylamine in spoiled fish, which is an indication of the presence of primary amines. R e a y ( 1 9 3 8 ) , Shewan ( 1 9 3 8 ) , R e a y and Shewan ( 1 9 4 9 ) , and S h e w a n ( 1 9 4 9 ) reported that during spoilage small amounts of mono- and dimethyl­

amine w e r e formed. B e a t t y and Collins ( 1 9 4 0 ) found that in spoiling cod muscle press juice some dimethylamine was produced, and that this occurred before any increase in trimethylamine b u t to a smaller extent than that of trimethylamine. V a r e l a and W o j c i e c h ( 1 9 5 6 ) and W o j c i e c h and V a r e l a ( 1 9 5 8 ) reported that the content of dimethylamine ranged from 0.2 mg. Ν per 100 g. (in fresh h a k e ) to 1.5 mg. Ν per 100 g.

(in spoiled h a k e ) . T h e rather large differences often found b e t w e e n the content of total volatile b a s e and that of T M A suggests that other intermediate amines could possibly b e present in addition to the am­

monia. T h e above data tend to substantiate this conclusion.

Steam heating for a short time produces insignificant quantities of formaldehyde from fish flesh, b u t heating under pressure produces measurable amounts of formaldehyde, together with dimethylamine, according to J a p a n e s e findings ( O t a , 1 9 5 8 b ) . T h e amount of D M A was almost proportional to the T M A O content in the flesh, whereas that of formaldehyde was not. T h e r m a l breakdown of T M A O was accelerated by the presence of fish extracts or certain kinds of amino acids. T h e degree of decomposition of T M A O in the presence of cysteine was proportional to the concentration of T M A O and cysteine. B u t the quantitative ratio b e t w e e n these two decomposition products was not the theoretical one. This might b e attributed to some subordinate reaction of formaldehyde with certain kinds of amino acids ( O t a , 1 9 5 8 b ) .

B . V O L A T I L E A C I D S

Next to the total volatile nitrogenous compounds, the content of total volatile or steam-volatile acids has probably received m o r e attention as a possible index of spoilage than any other group of substances. I n 1927

Tillmans et al. reported that the content of steam-volatile acids was not a reliable criterion of spoilage even though in certain cases it increased with spoilage. B o u r y ( 1 9 3 4 , 1 9 3 5 ) reported similar results for fish spoilage. T h e most consistent proponents of steam-volatile acids as a spoilage index during the past 2 0 years have b e e n Hillig and his co­

workers. Starting with the first paper b y Clark and Hillig ( 1 9 3 8 ) , a long series of results have b e e n presented on the value of the content of steam-volatile acids as an index of spoilage of r a w and canned tuna, mackerel, and sardines, of canned herring roe, and of raw cod: Hillig and Clark ( 1 9 3 8 ) ; Hillig ( 1 9 3 9 a , b ) ; Hillig and Knudsen ( 1 9 4 2 ) ; Hillig ( 1 9 4 4 , 1 9 4 5 ) ; Hillig et al. ( 1 9 5 0 a ) ; Hillig ( 1 9 5 4 , 1956a, 1 9 5 7 ) ; Hillig et al. ( 1 9 5 8 ) ; Hillig ( 1 9 5 8 ) . Malin ( 1 9 3 9 ) reported that the volatile acid n u m b e r of ether-extracted fatty acids could serve as an indicator of fish spoilage. Collins ( 1 9 4 1 ) found that the p r e s e n c e of acetic acid could b e used as a guide to the state o f preservation of cod, b u t that the determi­

nation was too time-consuming to b e used as a routine procedure.

C l a g u e ( 1 9 4 2 ) reported that for canned M a i n e sardines the content of volatile fatty acids was a useful indicator of spoilage, even though there was no correlation b e t w e e n the bacterial count and the content of volatile acids. B e a c h a m ( 1 9 4 6 ) studied the increase in total and individual volatile fatty acids in oysters and clams canned during spoilage. H e found that formic and acetic acids increased during spoilage, and that butyric acid was not present in good shellfish b u t was formed during spoilage.

Sigurdsson ( 1 9 4 7 ) concluded that above a storage temperature of 3 2 ° F . ( 0 ° C . ) the content of total volatile acids was a satisfactory measure of the condition of raw and canned herring. F ä r b e r ( 1 9 5 2 ) presented data showing that even though the content o f steam-volatile acids in­

creased somewhat during spoilage of raw tuna and mackerel, the increase was often small or variable, or else did not correlate well with sensory findings. In a series of canned tuna samples judged sensorily to b e in various states o f freshness or spoilage n o correlation was observed b e t w e e n the content of steam-volatile acids and the organoleptic judgments. Higasa ( 1 9 5 3 ) reported that the volatile acids increased to a greater extent during spoilage than did other criteria, such as p H or volatile bases. Suzuki ( 1 9 5 3 a ) reported that volatile acids could b e used to judge the freshness of fish. U c h i y a m a and Y o k o y a m a ( 1 9 5 3 b ) reported that for fish cakes the presence of propionic, butyric, and valeric acids could generally b e used as an index of spoilage, but that sometimes these volatile fatty acids were found even in fish cakes with no sensorily d e t e c t a b l e deterioration. T o m i y a m a and Yone ( 1 9 5 3 ) found no consistent values for the content of volatile acids at the onset of spoilage of

8 0 L I O N E L F Ä R B E R

"kamaboko" fish loaf. Luijpen ( 1 9 5 4 a , b ) reported that for salted herring there was a relation b e t w e e n the content of volatile acids and t h e state of spoilage. However, the practical application of the volatile acids content as a spoilage indicator was doubtful since volatile a c i d (e.g., a c e t i c ) , was added to m a n y herring preparations. Furthermore, the type of spoilage was influenced b y the salt content. Asakawa ( 1 9 5 3 , 1 9 5 4 , 1 9 5 7 ) , in a series of papers, described apparatus for the distillation of volatile acids; studied the volatile acid recovery in different amounts of distillate; studied the volatile acid production in different anatomical parts of fish and in different fish; proposed a numerical scale for the volatile acid content b a s e d on the ratio of the amount of acid in successive portions of distillate; and concluded that with advancing spoilage of fish flesh the higher volatile fatty acids b e c a m e m o r e prominent. F i e g e r and Friloux ( 1 9 5 4 ) found no relation b e t w e e n the volatile acids content and the early stages of shrimp spoilage.

Procedures for the determination of the content of volatile fatty acids have generally b e e n b a s e d on or are similar to those described b y Clark and Hillig ( 1 9 3 8 ) and b y F r i e d e m a n n ( 1 9 3 8 ) . Procedures for the chromatographic separation and identification of individual volatile fatty and organic acids have b e e n described b y R a m s e y and Patterson ( 1 9 4 5 , 1948a, b ) , R a m s e y and Hess ( 1 9 5 0 ) and van D a m e ( 1 9 5 7 ) . Vaisey ( 1 9 5 6 ) found that in nitrite-treated cod fillets the volatile acids content was too scattered, in relation to the sensory spoilage assessment, to serve as a useful index of spoilage. Orlandi ( 1 9 5 6 ) reported that volatile acids constituted an a c c e p t a b l e indicator of the degree of freshness in fish.

C. S U C C I N I C A C I D

Hillig and his co-workers (Hillig, 1 9 4 9 , 1954, 1956a, 1 9 5 9 ; Hillig et al., 1950b, 1 9 5 8 ) studied the production of succinic acid during spoilage of tuna and cod. F o r tuna it was suggested that the presence of succinic acid could serve as an index of spoilage, whereas for cod the content of succinic acid did not correlate with the state of spoilage.

D . I N D O L E AND S K A T O L E

Over the years various attempts have b e e n m a d e to use the content of indole and of its derivative methylindole or skatole as a measure of the state of freshness or spoilage in flesh foods. T e c h n i c a l procedures for its determination have b e e n studied by, a m o n g others, Fellers and Clough ( 1 9 2 5 ) and Chernoff ( 1 9 4 0 ) . T h e changes in the content of indole and skatole that occur in fish during spoilage h a v e b e e n reported b y Ottolenghi ( 1 9 1 3 ) , Guareschi ( 1 9 1 7 ) , Clough ( 1 9 2 2 ) , Tillmans and Otto

( 1 9 2 4 ) , Shostrom et al ( 1 9 2 4 ) , Pershin ( 1 9 3 5 ) , Smorodintzev and Diskina ( 1 9 3 6 ) , S h t e n b e r g et al ( 1 9 3 8 ) , K i n g et al. ( 1 9 4 5 ) , B e a c h a m

( 1 9 4 6 ) , D u g g a n and Strassburger ( 1 9 4 6 ) , D u g g a n ( 1 9 4 8 ) , A m a n o and T o m i y a ( 1 9 5 0 a ) , C a m p b e l l and W i l l i a m s ( 1 9 5 2 ) , F ä r b e r ( 1 9 5 2 ) , W i e r z h - chowski and Severin ( 1 9 5 3 ) , and B a r r y et al. ( 1 9 5 6 ) . T h e general conclusion that m a y b e drawn from these data is that w h e n e v e r indole and skatole w e r e present, the fish was organoleptically j u d g e d to b e definitely spoiled, b u t that during early stages of spoilage the a b s e n c e of indole and skatole could not b e used as a criterion of freshness. F o o d s that w e r e j u d g e d sensorily to show some spoilage w e r e not found to contain any or significant amounts of these substances.

E . H Y D R O G E N S U L F I D E

T h e situation with r e s p e c t to the determination of hydrogen sulfide as an index of spoilage is similar to that for indole and skatole. Various attempts have b e e n m a d e to use the content of hydrogen sulfide as a reliable m e a s u r e of t h e condition of fish and the results have b e e n variable. A m o n g these studies are those of E b e r ( 1 8 9 7 a , b ) , Fellers et al ( 1 9 2 4 ) , Almy ( 1 9 2 5 , 1 9 2 7 ) , B u d a g j a n ( 1 9 3 2 ) , Dobrovskii and Novikova ( 1 9 3 5 ) , Tilik ( 1 9 3 5 ) , Okolov and Shavskii ( 1 9 3 6 ) , B o u r y ( 1 9 3 7 ) , B o u r y and Schvinte ( 1 9 3 2 , 1 9 3 5 ) , Riddell et al ( 1 9 3 7 ) , T a n i k a w a ( 1 9 3 8 a ) , Stansby and L e m o n ( 1 9 4 1 ) , Sigurdsson ( 1 9 4 7 ) , and F ä r b e r ( 1 9 5 2 ) . Castell and G r e e n o u g h ( 1 9 5 7 b ) found that tetracycline antibiotics h a d no effect on the production of hydrogen sulfide b y spoilage bacteria.

Guareschi ( 1 9 1 7 ) reported the p r e s e n c e of mercaptans in spoiled fish.

T h e general conclusion from all the reported data is that wherever hydrogen sulfide was found, definite spoilage could b e established sensorily, b u t that spoilage could o c c u r with no hydrogen sulfide production.

F . C A R B O N Y L C O M P O U N D S

Among the possible products of bacterial action and of the c h e m i c a l breakdown of unsaturated fats are the carbonyl-containing substances, including keto acids, ketones, and aldehydes. D i a c e t y l , for example, is a well-known product of lactobacilli and certain lactic streptococci. M o n o - and dicarbonyl compounds have b e e n used as indicators of fat rancidity.

An attempt to use the content of carbonyl compounds reacting with bisulfite was m a d e m a n y years ago ( F ä r b e r , 1 9 5 2 ) . I t was found that the content of the bisulfite-binding substances increased with spoilage and varied directly with the sensory findings. I n c a n n e d fish, however, the correlation was not as close as for raw fish. Proctor et al ( 1 9 5 7 )

82 L I O N E L F A R B E R

found carbonyl compounds in haddock flesh b u t reported no difference b e t w e e n fresh and spoiled fish. Since m a n y of the carbonyl substances are odoriferous and could contribute to the sensory stimulation, further studies on their production m i g h t b e of value. T h e y probably contribute to and are included in the determination of total volatile substances discussed below.

G . S T E A M - V O L A T I L E O X I D I Z A B L E S U B S T A N C E S

As an approach to the c h e m i c a l determination of odoriferous com­

pounds formed during spoilage of foods, investigators have used the oxidizability of steam distillates as a measure of their condition. T h e y have compared the values thus obtained with sensory judgments. Among the oxidizing agents used w e r e acid potassium permanganate, acid potassium dichromate, and, less commonly, alkaline potassium perman­

ganate. T h e relation of the oxidizability of steam distillates to the state of freshness has b e e n investigated in various foods, b u t only a few studies refer to fish: Strohecker et al. ( 1 9 3 7 ) ; E r g o r o v a ( 1 9 3 9 ) ; T o m i y a s u et al.

( 1 9 5 2 ) ; Hillig et al, ( 1 9 5 8 ) . Holaday ( 1 9 3 9 ) and Hillig ( 1 9 5 8 ) have used the same general procedure with specific refinements to estimate the content of ethanol as a measure of spoilage.

This procedure yielded some interesting correlations with the sensory judgments and showed definite utility as a means of measuring spoilage of various foodstuffs. B u t the m e t h o d has certain disadvantages that limit its use: one, that a protein-free filtrate must b e prepared for the distil­

lation; another, that the steam distillation operation itself could con­

ceivably remove m a n y steam-volatile substances that are not essential to the odor. Heat-labile compounds m a y also b e broken down with the formation of steam-volatile products. T h e s e would then b e included in the evaluation of the total oxidizable substances. Finally, the physio­

logical act of smelling and the steam distillation of a markedly altered extract of the original sample are not strictly comparable operations.

H e n c e , the results of t h e two processes would not b e expected to agree very closely. T h e rather divergent results reported for steam-volatile oxidizable substances m a y b e due to these circumstances.

T o m i y a m a et al. ( 1 9 6 0 a ) avoided foaming b y removing the proteins through magnesium sulfate. T h e steam distillation was carried out in an alkaline reaction at p H 9.3. This m e t h o d gave a m o r e steeply mounting curve for the incipient spoilage of mackerel. T h e authors claimed that this modified form constituted the most sensitive w a y of measuring spoilage in m a c k e r e l flesh. A subsequent study confirmed these findings and it was maintained that this modified version o f the steam-oxidizable

substances m e t h o d gave a b e t t e r coverage of available volatiles ( T o m i ­ y a m a et al.9 1 9 6 0 b ) .

H . R E D U C I N G S U B S T A N C E S T H A T A R E V O L A T I L E A T R O O M T E M P E R A T U R E

A procedure that eliminated the objections enumerated a b o v e for the steam-volatile oxidizable substances was first reported b y L a n g et al.

( 1 9 4 4 , 1 9 4 5 ) . I n it, the sample tested was a press juice squeezed from either raw, processed, or c a n n e d fish. Air at room temperature was passed first through the sample juice, w h e r e it p i c k e d up and entrained any substances that w e r e volatile at a m b i e n t temperature, and then through a solution of potassium p e r m a n g a n a t e in normal sodium hydroxide. T h e amount o f p e r m a n g a n a t e r e d u c e d was used as the measure of t h e content of volatile material in the air w h i c h had passed through the sample. T h i s procedure was designed to duplicate chemically as closely as possible the physiological process of smelling or odor perception. I n b o t h processes air at a m b i e n t temperature is passed first over or through a sample; it then comes into c o n t a c t with the detecting mechanism, either the olfactory nerve endings or the alkaline solution of potassium perman­

ganate. I n b o t h processes the sample is not altered or changed, excepting that the soluble constituents are separated from the supporting structure for the c h e m i c a l detection process. T h e reagent chosen was found to b e the most sensitive oxidizing agent of a n u m b e r tested, including a c i d permanganate, acid dichromate, and eerie sulfate. I t r e a c t e d m o r e rapidly at room temperature than any of the other oxidants tested, and with a wide variety of organic and inorganic compounds that could b e e x p e c t e d to b e present in fresh or spoiled fish. T h e main exception was ammonia, which did not r e d u c e the reagent. T h e early exploratory studies indicated that the m e t h o d would give useful results and its further development was promising. S u b s e q u e n t reports b y F ä r b e r ( 1 9 4 9 , 1 9 5 2 ) and F ä r b e r and Cederquist ( 1 9 5 3 ) showed the applicability of the volatile reducing substances ( V R S ) procedure to a w i d e variety of odoriferous foods and food products in addition to fish. F ä r b e r and F e r r o ( 1 9 5 6 ) described a modification of the m e t h o d ( t h e air purification train was eliminated and a recirculating p u m p inserted to m a k e a self-contained u n i t ) , and showed the applicability of the modified m e t h o d to a w i d e variety of c a n n e d fish.

A review of this procedure as a useful index of spoilage of different kinds of raw and processed fish has b e e n published b y F ä r b e r and L e r k e ( 1 9 5 8 ) . S o m e data for the m e c h a n i s m of oxidation b y alkaline perman­

ganate solution have b e e n reported b y K a r e l et al. ( 1 9 5 7 ) . I t m a y b e of interest to point out that the V R S p r o c e d u r e for spoilage evaluation is essentially a relative one, the amount of reduction for any sample b e i n g compared to that for fresh samples of the s a m e species determined under

84 L I O N E L F A R B E R

the same conditions. T h e theoretically possible total reducing capacity and total recovery of all reducing oxidizable substances are therefore irrelevant concepts.

Reports of the usefulness of this m e t h o d as a spoilage indicator for marine and fresh-water fish have b e e n published b y Moorjani et al.

( 1 9 5 8 ) , b y Wittfogel ( 1 9 5 6 , 1958a, b ) , and b y W i t t f o g e l and G e b h a r d t ( 1 9 5 7 ) . R i e m a n n ( 1 9 5 2 ) reported on the content of total volatile nitrogen bases, trimethylamine, volatile reducing substances, and total bacterial count, as well as on the p H of cod fillets stored at 0 - 2 ° C . H e also compared their values with the organoleptic judgments on a numerical scale of 1-5, from very fresh to putrid. H e concluded that the determi­

nation of total volatile b a s i c nitrogen gave b e t t e r results than the other methods. A careful examination of the charts given, however, do not show the experimental basis for this conclusion. All the ranges for the various values at the corresponding sensory grades overlap and do not permit conclusions about the usefulness of the methods studied. Schmidt and M a y o h ( 1 9 5 5 ) reported data for some salmon spoilage in i c e and at 3 8 ° F . in air. No m a r k e d odors developed in ice and the increase in V R S was small, whereas in air there was a m o r e rapid spoilage with the emission of marked odors and a greater production of V R S .

C o m b i n i n g determinations V R S with that of T M A and of the percen­

tage of p i g m e n t e d b a c t e r i a before and after a 5-hour incubation period at 3 0 ° C . gave a useful indication of the degree of freshness. This allows a reasonable prediction of the storage capability or keeping quality of a raw fish sample ( F ä r b e r and L e r k e , 1 9 6 1 ) .

A m e t h o d for determining V R S b y microdiffusion analysis was devised b y Suzuki ( 1 9 5 9 b ) . W h e n comparing the V R S values of canned m a c k e r e l ( b o i l e d ) with those determined b y sensory judgment, a good correspondence was observed. I t was concluded that the freshness of the original raw fish can b e satisfactorily established from the V R S values of canned fish, b u t not from the amount of volatile b a s i c nitrogen

(Suzuki, 1 9 5 9 b ) .

Inasmuch as the V R S determination gives a measure of the total odoriferous substances present in a sample, it could b e expected that in general results from this m e t h o d would correlate with the sensory evaluation, which is also to a large extent b a s e d on the presence o f detectable odors. T h e studies carried out in the author's laboratory over the past ten years or more have led to the conclusion that determination of the content of volatile reducing substances is a useful and practical means to evaluate chemically the amount of spoilage in a wide variety of raw and c a n n e d fish and fish products.

I. H I S T A M I N E , H I S T A M I N E - L I K E S U B S T A N C E S , AND O T H E R S

I n r e c e n t years, interest in histamine as an index of spoilage of fish has b e e n revived. A special chapter ( V o l u m e I, C h a p t e r 1 0 ) is devoted to this intriguing subject with its m a n y evasive aspects. Several causal relationships still remain to b e elucidated. Until this has b e e n done, the usefulness of histamine as a spoilage indicator is limited. A few studies pertaining to its applicability for this purpose are available ( W i l l i a m s , 1 9 5 4 - 1 9 5 9 ; Hillig, 1 9 5 4 , 1956a; Yamanishi et al, 1 9 5 4 ; Torres-Acero F e r n a n d e z , 1 9 5 6 ) . A c h e m i c a l m e t h o d for the determination of histamine has b e e n described b y S a g e r and Horowitz ( 1 9 5 7 ) . E a r l i e r methods were of a biological nature, w h i c h r e d u c e d their dependability. O t a and K a n e k o ( 1 9 5 8 ) noted differences in histamine production b e t w e e n spoil­

ing fresh and t h a w e d muscles, even w h e n differences in deterioration b e t w e e n the two products, measured in amount of a m m o n i a and the mercuric chloride reaction, w e r e insignificant.

C o n c e r n i n g the possibility of histamine as a poisoning agent, see Volume I, C h a p t e r 10.

J . I O D I M E T R I C T I T R A T I O N AND I O D I N E U P T A K E

Truttwin ( 1 9 5 3 , 1954, 1 9 5 5 ) presented an iodimetric titration proce­

dure as a means to determine the extent of spoilage of various fish, including cod, hake, haddock, whiting, plaice, herring, and salmon. T h e originally r e c o m m e n d e d aqueous suspension was modified later to an aqueous a c e t o n e suspension of the fish. An iodine solution in potassium iodide was used as the titrant. L e v e l s for various grades of fish were also given. Orlandi ( 1 9 5 5 ) reported favorable results with this procedure in tests on seven major m a r i n e fishes. H e r e c o m m e n d e d as a freshness limit an iodine uptake of 7 ml. of 0.01 Ν iodine solution p e r gram of fish. T h i s method was also reported to have given good results b y d O r a z i o ( 1 9 5 6 ) . Varela and W o j c i e c h ( 1 9 5 6 ) found that there was an iodine uptake b y h a k e during spoilage b u t that the amount d e p e n d e d on the physical state of the fish and the extent of its comminution. T h e y c o n c l u d e d that this test was not a reliable spoilage indicator. W o j c i e c h and V a r e l a ( 1 9 5 8 ) found that the iodine uptake varied with the species of fish or shellfish.

( T h i s principle of iodine uptake as a spoilage indicator h a d b e e n tried on fish m a n y years ago with little or no success. A m o n g the previous reports on the subject are those b y Poluektov, 1 9 3 3 ; Okolov and Shavskii, 1936; and Shavskii and Vikoulov, 1 9 3 6 . )

T h e m e t h o d apparently is b a s e d upon the idea that during spoilage substances that r e a c t with iodine solution are formed. This premise is

8 6 L I O N E L F A R B E R

open to criticism, however, since an uptake of iodine is possible (e.g., b y proteins and fats) without any relation to spoilage development.

K . P R O T E I N H Y D R O L Y T I C P R O D U C T S

1. Amino, Carboxyl, and Sulfhydryl Group; Nonprotein Nitrogen, and Biuret Reaction

Over the years the contents of amino nitrogen and of nonprotein nitrogen have b e e n studied as spoilage indicators in fish with variable results. Clark and Alby ( 1 9 1 7 a , b ) studied the increase in amino nitrogen and found some increase with spoilage, b u t concluded that other more sensitive and reliable tests w e r e needed. B o k m a n ( 1 9 1 8 ) , van Driest ( 1 9 2 0 ) , Riff art ( 1 9 2 2 ) , Aleev et al ( 1 9 3 6 ) , and others h a v e reported a favorable correlation b e t w e e n the sensory judgment of freshness and the content of amino nitrogen. Nevertheless, the usefulness of the content of amino nitrogen or that of nonprotein nitrogen as a measure of early fish spoilage was questioned b y the majority of the reports, e.g., Tillmans and Otto ( 1 9 2 4 ) , B o u r y ( 1 9 3 4 ) , Boury and Schvinte ( 1 9 3 2 , 1 9 3 5 ) , Nickerson and Proctor ( 1 9 3 5 ) , Okolov and Shavskii ( 1 9 3 6 ) , Smorodintzev and Kruilova ( 1 9 3 6 ) , Salmon and L e G a l l ( 1 9 3 6 ) , Riddell et al ( 1 9 3 7 ) , Beatty and Collins ( 1 9 3 9 ) , van de V e l d e ( 1 9 4 0 ) , L a b a r r e and F o u g e r e ( 1 9 4 2 ) , Sigurdsson ( 1 9 4 7 ) , Partmann ( 1 9 5 1 ) , T a n i k a w a and Akiba ( 1 9 5 5 ) , T a n i k a w a et al ( 1 9 5 2 a , b , 1953a, b , c ) , Aim ( 1 9 5 6 ) , Vaisey ( 1 9 5 6 ) , Amano and Bito ( 1 9 5 1 ) , and Suyama and K o n o s a ( 1 9 5 7 ) .

Among the methods that have b e e n widely used to determine the content of amino groups are the formol titration of S0rensen, V a n Slyke's gasometric method, and the copper precipitation procedures of Pope and Stevens ( 1 9 3 9 ) . T h e biuret reaction in relation to spoilage determination has b e e n studied b y Proctor et al ( 1 9 5 7 ) , who concluded that it was of no significance for this purpose.

T h e reaction of ninhydrin with such compounds as amino acids and peptides, to give colored compounds, has b e e n used for studying protein- breakdown products during fish spoilage. F r e e amino acids and other ninhydrin-reactive substances in c o d muscle w e r e determined b y two- dimensional p a p e r chromatography to determine w h a t changes in the ninhydrin-reactive substances occur as the muscle deteriorates during refrigerated storage ( M i y a u c h i and Malins, 1 9 5 7 ) . I n ground samples stored for 6 days there occurred a decrease in the amounts of valine and of leucine. E x c e p t for these changes, n o other ninhydrin-reactive com­

pounds were found b y this experimental procedure.

E m p l o y i n g chromatographic procedures, R a n k e ( 1 9 6 0 ) noted several changes in the relative amount of free amino acids during

spoilage, some of w h i c h possibly could b e used for developing freshness tests ( s e e further V o l u m e I, C h a p t e r 1 6 ) .

T h e increase in sulfhydryl groups during spoilage has b e e n studied by Okolov and Shavskii ( 1 9 3 6 ) and b y Mori and H a t a ( 1 9 4 9 ) with divergent conclusions. T h e latter proposed the determination of sulf­

hydryl groups b y Anson and Mirsky's ferricyanide procedure as a method to estimate fish spoilage. T h e total sulfhydryl content of shrimp has b e e n followed during spoilage b y Kurtzman et al. ( 1 9 6 0 ) . T h e y found some increase in the cystine content of spoiled shrimp as well as in the total sulfhydryl content. Neither of these determinations, however, could b e considered a sensitive, reliable, and easily carried-out test for freshness or early spoilage.

R e l a t e d to the possible increase in amino groups during protein break­

down b y b a c t e r i a is the liberation of carboxyl groups. Tillmans and Otto ( 1 9 2 4 ) followed the increase in carboxyl groups b y the Willstätter and W a l d s c h m i d t - L e i t z double aqueous and alcohol titration procedure in a n u m b e r of fish species. T h e y found some increase in each case, even though it was rather small.

2. Tyrosine Value

T h e liberation of substances reacting like tyrosine with the phenol reagents of Folin and Denis ( 1 9 1 2 a , b ) and of F o l i n and Ciocalteu ( 1 9 2 7 ) and with their modifications has b e e n studied and proposed as a means to measure fish spoilage. T a r r and Bailey ( 1 9 3 9 ) reported that the tyrosine values for halibut "showed such a variation that it would b e difficult to establish a level above which fish could b e considered unfit."

Bradley and Bailey ( 1 9 4 0 ) m a d e a more extensive study of the m e t h o d for carp, herring, and salmon spoilage and r e c o m m e n d e d it as a useful indicator of the condition of the fish. Since this work a n u m b e r of other investigators have used the m e t h o d with varying results. W o o d et al.

( 1 9 4 2 ) , Sigurdsson ( 1 9 4 7 ) , Rasmussen ( 1 9 5 0 ) , Partmann ( 1 9 5 1 , 1954, 1 9 5 7 ) , Luijpen ( 1 9 5 4 a , b ) , V a r e l a and W o j c i e c h ( 1 9 5 6 ) , and W o j c i e c h and Varela ( 1 9 5 8 ) have reported data for the changes in tyrosine values during spoilage of cod, herring, hake, crab, and shrimp at various tem­

peratures. T h e consensus from all these observations is that the tyrosine value generally increases with advanced spoilage b u t that it is not very sensitive to changes occurring during the early stages of spoilage. This observation has also b e e n m a d e b y Soudan ( 1 9 5 0 ) in a review on spoil­

age and preservation of fish. At low temperatures the tyrosine formation is very insignificant even if spoilage b e c o m e s evident ( B r a d l e y and Bailey, 1 9 4 0 ; Luijpen, 1954a, 1 9 5 8 ) .

88 L I O N E L F A R B E R

Vaisey ( 1 9 5 6 ) reported that the increase in tyrosine value of cod fillets after a 3-hour incubation period at ambient temperature appeared to b e a m o r e satisfactory index of their condition than the original tyrosine value, w h i c h was quite variable. Basically this agrees with the findings of O t a and Ajisaka ( 1 9 5 3 ) , in several J a p a n e s e fishes, that the relative increase in amount of free tyrosine was, together with the changes in the ammonia values, most useful for the assessing of freshness of fish.

3. Mercuric Chloride precipitation

Amano and U c h i y a m a ( 1 9 4 8 , 1 9 4 9 ) , Amano et al. ( 1 9 4 9 ) , Amano and T o m i y a ( 1 9 5 0 a ) , and Amano ( 1 9 5 0 ) tried a freshness test described b y W a l k i e w i c z ( 1 9 3 6 ) employing mercuric chloride and reported some favorable results. Amano ( 1 9 5 4 ) stated that the test was used b y local health inspectors as a measure of spoilage. Buff a and Ambanelli ( 1 9 5 4 ) reported that for canned tuna and mackerel this test was useful for dis­

tinguishing slight from advanced spoilage b y the degree of turbidity and precipitation. T a n i k a w a et al. ( 1 9 5 2 a , b ) reported favorable results with mackerel b u t inconsistent results with fresh crab m e a t ( T a n i k a w a et al., 1 9 5 5 ) . I t was concluded to b e more of a qualitative than a quantitative method in studies on hake, squid, and shrimp ( V a r e l a and W o j c i e c h , 1956; W o j c i e c h and Varela, 1 9 5 8 ) .

L . N U C L E O T I D E S AND D E R I V A T I V E S

T h e importance of adenosinetriphosphate ( A T P ) and related nucleo­

tides and of sugar phosphates in tissue metabolism and muscle contrac­

tion has b e e n known for a long time. In recent years interest has increased in the content of these compounds in fish muscle and in their fate during rigor mortis, low temperture storage, and spoilage. T h e relation of the nucleotides and their derivatives to the flavor of fish flesh and its changes during storage and to the development of discolorations during heat processing has also b e e n pointed out in the last few years ( s e e Jones, 1 9 6 1 ) . F r o m the work reported to date on the nucleotides and their changes during storage (including that of Shewan and Jones, 1957;

Saito et al, 1 9 5 9 ; Jones and Murray, 1 9 6 0 , 1 9 6 2 ; Burt, 1 9 6 1 ; Bito and Amano, 1 9 6 2 ; Tomlinson and Geiger, 1 9 6 2 ; and Kassemsarn et al, 1 9 6 3 ) it appears that A T P breaks down rapidly during the early stages o f storage with the formation of such derivatives as adenylic acid, inosinic acid, inosine and hypoxanthine, and ribose. I t has b e e n suggested that the content of hypoxanthine m a y b e a useful index of quality and fresh­

ness of fish muscle during the early stages of chill storage, before bacterial breakdown of the constituents b e c o m e s predominant and important.