A. G E N E R A L
Well-directed and continuous surveillance o f the quality of the final product is always called for. Such quality control can b e effected either through process control, as advocated b y Dreosti and R o w a n ( 1 9 5 6 ) and R i e m a n n ( 1 9 5 7 ) , or b y regular examination of the product organo-leptically, chemically, and microbiologically.
Inspection services have b e e n maintained for years in most countries and these may, besides carrying out examinations, impose or offer certain rules for the canning of various products. S u c h rules are not compulsory, b u t usually canners h a v e found it to their advantage to comply with them ( J u l , 1 9 5 1 ) .
In addition, e a c h canner will of course regularly examine his own products and b e interested in finding the most efficient w a y to do this with the usually somewhat limited e q u i p m e n t at his disposal. T h e use of a standard form of report will ensure an orderly examination proce
dure and neat recording of findings, and facilitate interpretation and com
parison of results. S u c h a standard form has b e e n proposed b y Kefford ( 1 9 5 3 ) . T h e entire procedure is treated in great detail in an authoritative booklet ( C h e f t e l , 1 9 5 7 ) .
A growing willingness to cooperate can b e noted on the part of can
ners in the most industrialized countries. This, and an increasing openness toward the competitor, will p a v e the w a y to joint enterprises in the examination of the c o m m o n product. Thus, 1 9 5 1 witnessed the first
"tuna cutting" w h e r e tuna canners from all over California judged e a c h other's products under code (Anonymous, 1 9 5 2 d ) . In 1 9 5 6 , a petition was filed b y the U. S. National Canners Association proposing standards for canned tuna (Anonymous, 1 9 5 6 d ) . Such direct interest in quality control is likely to raise the general level of the quality of canned seafoods.
B . S A M P L I N G
T h e question of sample size has b e e n a matter of m u c h debate. I t is essentially a question of howT m a n y cans from a certain production
must b e examined in order to detect a type of spoilage, w h i c h is or can b e m a d e manifest in only a certain p e r c e n t a g e of the total n u m b e r of cans.
R i e m a n n ( 1 9 5 7 ) quotes the following figures for the p e r c e n t a g e o f spoiled cans remaining u n d e t e c t e d with a probability o f 5 % ( P = 0 . 0 5 ) b y examining increasing numbers of cans:
Number of cans Number of spoiled examined cans not detected (%)
5 52 10 31 20 17 50 7 100 3.5 200 1.8 500 0.7 1000 0.4 T h u s some 3 0 0 cans would have to b e examined to detect a spoilage
level o f 1 % , and even then there is a c h a n c e of 5 % that it would escape being noticed. W h i l e this situation is essentially similar for all kinds of spoilage, microbial or otherwise, R i e m a n n points out that the n u m b e r of cans to b e examined bacteriologically would have to b e about doubled in order to obtain the same detection efficiency, b e c a u s e only about half the nonstable cans spoil during a reasonable incubation time.
F o r all practical purposes the examination o f such a n u m b e r of cans is impossible, and usually m u c h smaller numbers are taken. M u c h will depend on the purpose for w h i c h the sampling is done. Routine inspec
tion for product quality will as a rule b e b a s e d on a small percentage of the production, whereas efficiency of processing will h a v e to b e tested b y incubation of m u c h larger numbers of cans. Kefford indicates ( 1 9 5 3 ) that at least 3 cans are desirable as a sample for general examination. However, the sample usually comprises 12 spoiled and 12 normal cans from the same b a t c h in one examination (Kefford and Murrell, 1 9 5 5 ) . Cheftel ( 1 9 5 7 ) considers 6 cans showing defects and 6 control cans to b e a minimum for examination.
T h e C a n n e d F i s h Inspection L a b o r a t o r y of the Canadian D e p a r t m e n t of Fisheries has, for routine inspection, withdrawn cases from production parcels according to t h e following rule:
Number of cases in parcel Minimum number of cases withdrawn
25 3
26-50 6 51-100 12 101-500 18 501-1000 24 1001-5000 48
w h e r e b y one can is taken from every case. This can is opened after determination of vacuum, and the contents examined for quality. Only if the latter findings give reason for doubt are further cans examined; if suspected to b e underprocessed, they are incubated (Anonymous, 1 9 5 2 g ) .
B l o w n cans or cans known to b e spoiled without swelling are examined without previous incubation; other cans are incubated as set forth below. Careful registration should b e m a d e of all particulars of the sample: origin, number, size, etc., and of all details of t h e canning process. As said before, it is useful to have these data registered in a standard form so that no details are overlooked and comparisons with other samples are facilitated.
C. D I R E C T E X A M I N A T I O N
Examination of cans begins with a careful description of the outside appearance, for which the following terms are conventional (Kefford, 1 9 5 3 ) :
Flat can: a can on which both ends are flat or concave.
Flipper or springer: a can on w h i c h one e n d is b u l g e d and, w h e n pressed, causes the other end to bulge, or springs out again itself; the term springer is also applied to rectangular tapered cans in which the springiness is usually more evident in the sides than in the ends.
Soft swell: a can on which both ends are bulged b u t yield to moderate pressure.
Hard swell: a can on w h i c h b o t h ends are bulged and unyielding.
V e r y advanced swells m a y show p e r m a n e n t distortion and are described as buckled.
Leaker: a can showing visible leakage of the contents through the seams, or through perforations or nail holes.
O t h e r deformations are:
Dents: m e c h a n i c a l injuries sufficiently pronounced to cause significant reduction in the internal volume of the can, or to deform the seams.
Panels: flat vertical dents observed only on the larger-size cans and due to partial collapse of the can b o d y under high internal vacuum.
Palings: narrow vertical flats around the body of the can caused b y fabrication against the "grain" of t h e tin plate without flexing.
Peaks: pyramidal deformities of the ends near the double seams caused b y permanent strain during retorting or cooling.
Perforations: points where corrosive attack, either internal or external, has b e e n sufficiently strong to eat through the tin plate.
Nail holes: punctures caused b y c a s e nails.
After the can is w a s h e d and weighed, before opening, it is decided whether aseptic sampling of the contents must p r e c e d e the examination
for quality, or w h e t h e r v a c u u m must b e tested or head-space gas analyzed.
E a c h of these practices will b e dealt with below.
It is useful, in direct examination, to open the can lid in such a w a y that the seams are left untouched for later examination and testing ( K e f ford and Murrell, 1 9 5 5 ) . F o r routine examination o f the product, the odor, color, and firmness are tested and carefully recorded, as these findings can serve as important supplementary information in cases w h e r e bacterio
logical investigation does not give clear-cut results. T h e odor o f t h e product, wnen judged b y an experienced person, is still considered the most sensitive spoilage test.
B y w a y of illustration: routine examination of tuna as p r a c t i c e d b y the Californian F i s h Canners Association involves (Anonymous, 1 9 5 2 d ) net and drained or pressed w e i g h t of the fish per can, m e a s u r e m e n t of color immediately upon opening and after the fish is uniformly ground, rating of p e r c e n t a g e of flakes in solid pack, sizing of chunk and grading in relation to the set standards for chunk pack, v a c u u m test, and indica
tion o f oil quality. F o r salmon, a grading executed b y the D e p a r t m e n t of Fisheries L a b o r a t o r y o f C a n a d a is b a s e d on: n e t weight, vacuum, firmness of the flesh as measured b y a penetrometer ( s e e Craven and Dassow, 1 9 5 2 ) , color as measured b y a colorimeter, free liquid and free oil, and an organoleptic j u d g m e n t of freshness. In case of doubt about the latter, the content of C 02 in the salmon m e a t is determined chemically ( B o l t o n , 1 9 5 2 ) .
E s p e c i a l l y when bacteriological tests are negative, further c h e m i c a l analyses can b e helpful in determining t h e cause of spoilage. I n blown cans, the analysis of head-space gas m a y offer a major indication. I n cases of decoloration, t h e search for heavy metals is warranted. M o s t chemical tests, however, must b e standardized for e a c h individual product.
D . B A C T E R I O L O G I C A L E X A M I N A T I O N
This line of investigation serves to answer a n u m b e r of questions, e.g.:
( 1 ) F o r cans in w h i c h organoleptically no spoilage is observed:
( a ) W h e t h e r the contents a r e commercially sterile, i.e., of normal keeping quality, through the incubation test.
( b ) W h e t h e r the contents are fit for h u m a n consumption.
( 2 ) F o r spoiled cans:
( a ) T h e cause o f spoilage, w h e t h e r due to microbial d e c a y before canning (spoiled r a w m a t e r i a l ) .
( b ) W h e t h e r due to underprocessing or to contamination after processing ( l e a k i n g c a n ) .
T o answer the latter two questions, the n u m b e r and kind of viable
bacteria in the contents are sought b y subculturing, and the total n u m b e r of ( d e a d ) b a c t e r i a b y direct microscopic count.
1. Incubation
Considering that c a n n e d fish products normally offer excellent b a c teriological media which, as a rule, do not contain bacteriostatic com
pounds, incubation of cans at a temperature above normal is still the most convenient test for the presence of viable bacteria. T h e r e is some variation of opinion, however, about the temperature and duration of incubation to b e preferred.
I t is usual to keep cans at two different temperatures:
(a) 5 0 - 5 5 ° C . ( 1 2 2 - 1 3 1 ° F . ) . This range is used to test for obligate thermophiles (Cheftel, 1 9 5 7 ) . As these bacteria show little tendency to remain dormant, the incubation period can b e kept short. Kefford and Murrell ( 1 9 5 5 ) , incubating at 5 0 ° C , advise 1-2 weeks, b u t it seems quite feasible to shorten the test to some 4 days b y increasing the incubation temperature to 5 5 ° C . ( R i e m a n n , 1 9 5 7 ) .
(b) 3 0 - 3 7 ° C . ( 8 6 - 9 9 ° F . ) . In this temperature range, practices differ more widely. R i e m a n n ( 1 9 5 7 ) names 3 7 ° C . as t h e most common tem
perature for mesophiles b u t at the same time cites experimental findings of a greater rate of swelling at 3 0 ° C , a temperature used b y a great num
ber of laboratories (Scott, 1 9 5 3 ) . Among the mesophiles, germina
tion of spores is influenced b y the large variety of factors prevailing in
side the can, such as composition of the food, presence of C 02, the p H , heat activation, etc. Multiplication and toxin production of Cl. botulinum in spoiling fish products increase with a dropping redox potential ( A n d o and Inoue, 1 9 5 8 ) .
In certain media, e.g., in the presence of fat, germination can b e m u c h retarded, so that very long incubation times are needed to cause all nonsterile cans to swell. R i e m a n n cites an experiment in which the number of cans (inoculated with Cl. sporogenes and then given 8 0 - 1 0 0 % processing) after 2-weeks incubation was only 5 6 % of the n u m b e r of cans blown after 12 weeks. In practice, it is seldom possible to keep cans for more than a few weeks; it is best to keep them as long as possible, to mention the results only in connection with incubation time, and to use the outcome of medium temperature incubation tests with some caution.
Rules for incubation temperatures for semipreserves are m u c h more complicated due to the varying p H and amounts of bacteriostatics added to these products. T h e treatment of these falls outside the subject of this chapter.
2. Culturing
A m o r e direct w a y to search for viable b a c t e r i a in can contents is to take a sample aseptically and inoculate a n u m b e r of suitable media.
This manner of investigation is usually more rapid than incubation and allows at least certain main categories of microorganisms to b e distin
guished. On the other hand, the mass represented b y the sample is usually very small in relation to total volume of the can, which will show up with far greater certainty in incubation.
a. A S E P T I C S A M P L I N G
D i r t y cans are first c l e a n e d with soap a n d w a t e r or with w a r m quaternary ammonium compound solution. T h e top of the can is then dried most conveniently b y wiping it with ethanol; burning the remaining alcohol is, however, not sufficient to sterilize the surface, w h i c h must b e flamed.
Before the can is opened, it is decided from t h e previous quality ex
amination whether head-space gas must b e analyzed ( s e e b e l o w ) . I f no gas analysis is required, the can lid can b e p i e r c e d b y a sharp pointed steel rod. O f course this is also flamed beforehand. A further precaution is to prevent air contamination b y attaching a shield to the rod. Cheftel ( 1 9 5 7 ) describes a m e t h o d for puncturing the can through a sterile glass funnel in order to avoid spouting of the contents. Absorbent cotton soaked in 1 % sodium hypochlorite solution can also prevent contamination
( R i e m a n n , 1 9 5 7 ) .
A sample of the contents is then taken b y using a sharpened steel spatula and forceps. B o t h are easy to sterilize b y flaming b e t w e e n e a c h two cans. Some, however, prefer to use cotton-plugged cork borers; a n e w sterile one has then to b e taken for each can. In solid p a c k cans, the sample should b e taken from the center as well as from the surface.
L i q u i d is sampled b y sterile pipettes. F o r all manipulations in a routine laboratory, the use of ultraviolet lamps m a y b e a help in preventing air contamination.
T h e solid samples measuring some 1-2 g. can b e diluted b y suspension and then various m e d i a can b e inoculated.
b . C U L T U R E M E D I A
T h e largely empirical nature of bacteriological culturing has resulted in an a b u n d a n c e of recipes for culture media. I t is not in the scope of this chapter to quote such recipes in detail; for these the reader will find a wealth of information in B a u m g a r t n e r ( 1 9 5 6 ) and T a n n e r ( 1 9 4 4 ) on the microbiology of canned foods in general or in Cheftel's ( 1 9 5 7 ) compre
hensive work on the subject in fishery products.
O n the whole, all media supporting good growth under either aerobic or anaerobic conditions will b e useful. As stated earlier, however, it must b e kept in mind that the germination of dormant spores is influenced b y many factors: for Clostridia, the lag phase seems to depend a great deal on redox potential, b e i n g zero only below E H = — 4 5 m V ; for heat-treated anaerobes, germination is said to b e favored b y the presence of soluble starch, that of other anaerobes b y N a H C 03 ( s e e Riemann, 1 9 5 7 ) .
Normally a dextrose-tryptone broth is incubated at 3 7 ° C . and 5 5 ° C . for aerobes, and a liver broth at the same temperatures for anaerobes ( B a u m g a r t n e r , 1 9 5 6 ) ; for the latter, anaerobic conditions can most con
veniently b e r e a c h e d b y adding m i n c e d liver to the tubes and steaming for 2 0 min. before inoculating, while m a i n t e n a n c e of anaerobic condi
tions is ensured b y topping the medium with a paraffin or agar seal. A deep brain medium, to which neopeptone has b e e n added, is a very suit
able routine medium ( R i e m a n n , 1 9 5 7 ) . T h e brain should form a 5-cm layer in the tubes and is inoculated deep. R e d o x potential is b e l o w — 2 0 mV; p H is not easily upset and the medium permits sporulation and growth of most b a c t e r i a found in canned fish. Since obligate thermophiles are seldom found, incubation can b e restricted to 4 days at 3 7 ° C .
F o r further investigation of special cases of spoilage, special m e d i a m a y b e employed, e.g., dextrose-tryptone agar for "flat sour" organisms, sulfite agar for H2S - f o r m i n g thermophilic anaerobes, and a corn-liver medium for sacchrolytic thermophilic anaerobes.
A petri dish poured with inoculated nutrient blood agar and incubated at 2 0 ° C . for 4 - 5 days is very convenient for detecting non-spore-formers.
After incubation these plates are inspected visually and microscopically and, i f necessary, a further estimation of the n u m b e r of viable bacteria present can b e obtained b y inoculation with 1:10 or 1:100 dilutions of the original sample.
Aerobic spore-formers m a y then b e detected b y making streaks from the original cultures (e.g. brain m e d i u m ) after it has b e e n pasteurized at 8 0 ° C . for 2 0 min. anaerobes b y subculturing in a variety of milk m e d i a as described b y R i e m a n n ( 1 9 5 7 ) or, perhaps more specifically, in the medium described b y Mossel et al. ( 1 9 5 6 ) . T h e milk m e d i a can b e ren
dered inhibitory to Bacillus species b y the addition of sodium azide ( N a N8) .
F o r the normal nonacid or low-acid canned fishery products, it is generally agreed that even mild processing will invariably kill all non-spore-forming microorganisms. W h e n e v e r non-spore-formers are found in canned fish, they can have b e e n introduced only after processing, through leaks in the can. Consequently, as R i e m a n n points out, the counting of anaerobic spore-formers will seldom b e troubled b y the presence of
non-spore-formers, as underprocessing and can l e a k a g e are two essentially independent aberrations.
3. Microscopic Examination
Bacteriological examination is completed b y a microscopic search for bacteria, in spoiled cans directly upon sampling or after incubation.
Usually a w e t preparation of smeared contents examined under phase contrast microscope is sufficient to detect microorganisms. Otherwise, smears can b e stained with carbol-fuchsin or crystal violet. T h e presence of one or more b a c t e r i a p e r m i c r o s c o p e field warrants further examination ( R i e m a n n , 1 9 5 7 ) ; Kefford ( 1 9 5 5 ) considers 1 5 b a c t e r i a per field a suffi
cient n u m b e r to cause spoilage.
If, in such cases, no growth is obtained b y culturing, this m a y point to preprocessing spoilage or else to autosterilization, although in the latter instance it is often difficult to detect cells.
F u r t h e r differentiation of t h e cells b y the microscope can help to indicate whether culturing is necessary and w h a t m e d i a are most likely to b e successful.
E . E X A M I N I N G C A N AND F I L L I N G
1. Head-space Depth and Volume
In canning fish products, as with m a n y other foods, it is useful to leave a certain amount of head-space while filling the can, to allow for expansion of the contents during processing. T h e amount of head-space varies with the kinds of product and processing and has largely to b e found out empirically. O n c e a standard is set, filling should b e c h e c k e d according to these rules either before closing the can or, more exception
ally, along with the examination of the canned product as a whole.
Head-space depth is very easily determined, e.g., b y placing a b a r across the top of the opened can, w h i c h is provided with a series of prongs of increasing length ( 2 / 1 6 to 1 0 / 1 6 in. or 3 . 2 — 1 6 mm., in steps of 1 / 1 6 in.) and pointing downwards. Head-space depth is read from the prong that just touches the liquid surface. Besides this simple instrument devised b y the F i s h Industry R e s e a r c h Institute of C a p e T o w n , Kefford ( 1 9 5 4 b ) mentions t h e head-space depth gauge used b y C a m p d e n R e search Station, E n g l a n d . This consists of a graduated disk with a milled edge and provided with a tapering slot which engages a pin on a flat rod.
W h e n the leveled tip of the rod meets its reflection in the liquid surface, the pointer indicates the head-space depth in units of 1 / 1 6 in.
T h e s e simple measurements of head-space depth are convenient for routine inspection and control purposes, b u t will seldom b e applicable to solid packs such as most fish products are; moreover, they give
insuffi-ciently accurate data for experimental work. F o r this, head-space volume measurement is m u c h m o r e useful. Mitchell ( c i t e d in Kefford, 1 9 5 4 b ) devised a method using the Nicholson hydrometer. T h e measurement con
sists of weighing the filled and closed can in air and in water. F r o m these weights, the internal capacity of the can and the volume of the contents can b e calculated; the difference b e t w e e n these two quantities is the head-space volume. T h e article cited contains a detailed prescription for the construction of a hydrometer for this purpose.
2. Vacuum
C a n vacuum is expressed as the difference (in inches or mm. of H g ) b e t w e e n internal pressure in the can and atmospheric pressure. V a c u u m is in the range of 1 5 - 2 5 in. ( 3 8 0 - 6 3 5 m m . ) Hg. In commercial canning practice, 7 in. ( 1 8 0 m m . ) H g is commonly a c c e p t e d as a minimum sat
isfactory vacuum. T h e South African B u r e a u of Standards ( 1 9 5 1 ) has in
corporated in its specifications for most canned fish a requirement that the minimum vacuum shall b e 5 in. ( 1 2 5 m m . ) H g at 2 4 ° C . ( 7 5 ° F . ) .
T h e accurate measurement of vacuum b y means of mercury manom
eter is rather complicated. M o r e simple is the indirect tapping method widely used in Australian canneries for checking stacks prior to casing and shipment (Kefford, 1 9 5 4 a ) . O n e end of the can can also b e subjected to increasing vacuum, until it "flips" w h e n the internal pressure exceeds that of the exterior. I n t h e spherometer test, internal v a c u u m is estimated from the depth of the concavity of the lid. Shiga and K i m u r a ( 1 9 5 3 ) measure the size of a light disk reflected from the c o n c a v e can ends from a small electric lamp. Another way to estimate v a c u u m without destroy
ing the can was recently introduced (Shiga, 1 9 5 9 ) b y making use of the "Moire phenomenon," i.e., the appearance of rings on a disk of glass plane grating which is placed on the flat cover of a can without ex
pansion rings. I n all these crude methods of v a c u u m determination, factors such as diameter of the can end, stiffness of the tin plate, and pat
tern of the expansion rings m a y greatly influence the results. T h e s e meth
ods are therefore only applicable to standard production and for routine control.
T h e conventional w a y of testing can vacuum consists of applying a small dial gauge of the Bourdon type fitted with a puncturing needle and a soft r u b b e r gasket to the lid o f the can. V a c u u m determination as read from the dial after puncturing the can lid is subject to errors from various sources. S o m e of these are errors in reading, due to the design and method of graduation of the gauges. Errors in calibration of t h e reader of 1-1.5 in. ( 2 5 - 3 8 m m . ) H g are common and it is advised b y Kefford ( 1 9 5 4 a ) to calibrate gauges periodically against a mercury ma
nometer. T h e most obvious source of error is, however, the internal
vol-u m e of the Bovol-urdon tvol-ube and its connections, w h i c h m a y b e approx
imately 3.5 ml. W h e n this volume, containing gas at atmospheric pressure, is connected with the head-space of a can, the recorded vacuum is less than the true original v a c u u m in the can, the more so the smaller t h e head-space volume and the lower its pressure. Thus errors up to 3 3 % have b e e n noted.
T h e F I R A vacuum gauge, developed b y the British F o o d Manufac
turing Industries R e s e a r c h Association, obviates this most important
turing Industries R e s e a r c h Association, obviates this most important