Quick Pick By Commodity / Key Topic Animal health Biotechnology Dairy Eggs Employment Opportunities Feeds Fertilizers Fish and Seafood Food Recalls Forestry Fresh Fruits and Vegetables Grains Honey Horticulture Meat Hygiene Pet Imports Plant Biosafety Plant Breeder's Rights Potatoes Processed Products Publications Retail Food/Labelling Seeds Variety Registration Veterinary Biologics

Main Page - Animal Products Main Page - Fish, Seafood and Production Acts & Regulations Canadian Shellfish Sanitation Program Import Inspection Program - Import Alert List Export Information Quality Management Program Product Inspection - Fish List - Questions & Answers Fish Inspection Manuals Communiques/ Industry Notices Fish FAQ Related Sites Offices

iii

TABLE OF CONTENTS

iv

LIST of TABLES

Table 1. Weight (Kg) of pink, sockeye, and chum salmon used in each of the studies.

Table 2. Sampling schedule for the canning of the three species of salmon held under ambient and iced storage conditions.

Table 3. List of can codes used for the three species.

Table 4. Results of the Sockeye Salmon Sensory Set at the November 1994 Orientation Workshop.

Table 5. Results of the Pink Salmon Sensory Set at the November 1994 Orientation Workshop.

Table 6. Results of the Chum (keta) Salmon Sensory Set at the November 1994 Orientation Workshop (K= Fraser River chum, KX= Qualicum River chum).

Table 7. Accept and reject scores for sensory evaluation of canned sockeye salmon spoilage sets by eight DFO Inspectors during Internal DFO canned salmon workshop January 1995.

Table 8. Accept and reject scores for sensory evaluation of canned pink salmon spoilage sets by eight DFO Inspectors during Internal DFO canned salmon workshop January 1995.

Table 9. Accept and reject scores for sensory evaluation of canned chum salmon spoilage sets by eight DFO Inspectors during Internal DFO canned salmon workshop January 1995.

v

LIST of FIGURES

Figure 1. Comparison between the change in temperature within a tote full of Fraser River chum salmon held at ambient temperature and the actual room temperature where the totes were held.

Figure 2. Comparison between the change in abdominal temperature of Fraser River chum salmon held in slush ice or at ambient temperature.

Figure 3. Comparison between the average change in abdominal temperature over time of pink and sockeye salmon held in slush ice or at ambient temperature.

Figure 4. TVB-N concentration in sockeye salmon steaks from fish which had been held at an ambient temperature of 14C for 0-120 hours.

Figure 5. TVB-N concentration in sockeye salmon steaks from fish which had been held in slush ice for 0-19 days.

Figure 6. Median ethanol (EtOH) , cadaverine (Cad), and putrescine (Put), concentrations from canned sockeye salmon which had been held at ambient temperature or in slush ice for varying periods prior to canning. SA= sockeye ambient (intervals in hours), SI= sockeye ice (intervals in days).

Figure 7. TVB-N in pink salmon held at ambient temperature

Figure 8. TVB-N in pink salmon held in slush ice

Figure 9. Median ethanol, cadaverine and putrescine in pink salmon

Figure 10. Median ethanol, cadaverine and putrescine in Qualicum chum salmon

Figure 11. Mean DFO expert analyst line-scores versus mean line scores for industry analysts

Figure 12. Mean DFO expert analyst line-scores versus industry analyst 8

Figure 13. Line-score values for DFO analysts and the mean for each sockeye salmon spoilage increment as observed under non-blind conditions.

Figure 14. Line-score values for DFO analysts and the mean for each pink salmon spoilage increment as observed under non-blind conditions.

Figure 15. Line-score values for DFO analysts and the mean for each Fraser River chum salmon spoilage increment as observed under non-blind conditions.

Figure 16. Line-score values for DFO analysts and the mean for each Qualicum River chum salmon spoilage increment as observed under non-blind conditions

Figure 17. Line-score range for each sample unit examined in the demonstration set for sockeye salmon, (DFO analysts indicated by diamonds, and industry analysts indicated by solid bars).

Figure 18. Line-score range for each sample unit examined in the demonstration set for pink salmon, (DFO analysts indicated by diamonds, and industry analysts indicated by solid bars).

Figure 19. Line-score range for each sample unit examined in the demonstration set for Fraser River chum salmon, (DFO analysts indicated by diamonds, and industry analysts indicated by solid bars).

Figure 20. Line-score range for each sample unit examined in the demonstration set for Qualicum River chum salmon, (DFO analysts indicated by diamonds, and industry analysts indicated by solid bars).

vii

LIST of PARTICIPANTS

Susan Schenkeveld, Canadian Food Inspection Agency *
Andrew Chan, Canadian Food Inspection Agency *
Preston Chan, Canadian Food Inspection Agency *
Clive Cosham, Canadian Food Inspection Agency *
Nigel Harrison, Canadian Food Inspection Agency *
Debbie Koo, Canadian Food Inspection Agency *
Jennifer Lee, Canadian Food Inspection Agency *
David McLachlan, Canadian Food Inspection Agency *
Rebecca Reid, Department of Fisheries & Oceans
Greg Riddell, Canadian Food Inspection Agency *
Klaus Schallie, Canadian Food Inspection Agency *
Kevin Cheung, , Canadian Food Inspection Agency *
Steve James, Canadian Food Inspection Agency *
Anthea Fok, Canadian Food Inspection Agency *
Paul Wheeler, Canadian Food Inspection Agency *
Steve Trueman, formerly with the Department of Fisheries & Oceans
Jane Reid, formerly with the Department of Fisheries & Oceans
Laura MacDonald, formerly with the Department of Fisheries & Oceans
Roger Gibb, B.C. Packers Ltd.
David Yao, B.C. Packers Ltd.
David Calkins, Calkins & Burke Limited
Ralph Drew, Canadian Fishing Company
Alnoor Hirji, Canadian Fishing Company
Ross Young, Canadian Fishing Company
Terry Babcock, Coastwise Fisheries
Don Millerd, Great Northern Packing
Karen Damkjar, Great Northern Packing
Michel Benoit, J.S. McMillan
Satinder Sanghera, J.S. McMillan
Douglas Safarik, Ocean Fisheries Ltd.
Steve Yip, Ocean Fisheries Ltd.
Alison France, Seafood Products
Rob Paterson, Seafood Products
Brian Dougall, Shafer Haggart Limited
Gwen Hunter, St. Jean’s Cannery
Nirmal Sandhu, St. Jean’s Cannery
* formerly Department of Fisheries & Oceans, Inspection Branch

viii

ABSTRACT

P. Chan, D. Koo, D. McLachlan, R. Reid, K. Schallie and S. Schenkeveld 1999. Controlled spoilage of Pacific salmon for use as an indicator of decomposition in the canned product. Tech. Rept. Canadian Food Inspection Agency No. P0088E-99 44 p.

The goal of this study was to prepare sets of canned salmon for chemical and organoleptic evaluation from salmon that had been allowed to spoil under controlled conditions. These sets were then used by government and industry analysts for training in the determination of whether specific samples of canned salmon were acceptable or should be rejected for sale in the Canadian market. Sockeye (Oncorhynchus nerka), pink (Oncorhynchus gorbuscha) and chum (Oncorhynchus keta) salmon captured in the commercial fishery were canned using commercial equipment after having been held on ice or at ambient temperature for varying periods of time. Samples of the canned salmon were analysed for ethanol, cadaverine and putrescine. Total volatile base nitrogen (TVB-N) was quantified in samples of salmon taken from the various spoilage sets prior to canning. All chemical parameters increased in concentration during storage. Increases occurred more rapidly at ambient temperature than on ice. Ethanol and cadaverine rose more rapidly than putrescine and total volatile base nitrogen. Sensory evaluation by government and industry analysts of samples of canned salmon over the full range of storage conditions provided a consensus on the point at which each of the three species of salmon, canned after storage, crosses the line from acceptable to not acceptable. Thus, sockeye salmon was acceptable to most analysts after storage for up to 48 hours at ambient temperature and up to 5 days on ice. These chemically and organoleptically characterised sets of canned salmon provide a valuable tool for the instruction and evaluation of industry analysts.

ix

RESUME

1.0 Introduction

The importance of developing sensory evaluation techniques for Canadian fish products has been recognised for some time (York 1989). Furthermore, the international harmonisation of sensory evaluation techniques is important for international trade in foods of aquatic origin (Rielly and York 1994, Griffin 1998).

During the summer and fall of 1994, Inspection branch of DFO and the members of the British Columbia (BC) fishing industry conducted a joint study in which three species of Pacific salmon, sockeye, pink, and chum, were spoiled in a controlled experiment. The object of the study was to determine the point at which decomposition in canned salmon would be rejected by a trained sensory panel. All fish begin a gradual process of decomposition as soon as they are harvested. Although this process cannot be avoided, prompt and careful handling can reduce its rate. By producing a series of cans of salmon, which have undergone controlled spoilage, an incremental increasing level of decomposition will result. Sensory panels and chemical analysis can then be used to monitor the gradual change in the fish and determine the point at which the product becomes unacceptable for human consumption. This information can be used to judge future production, and determine whether particular lots of salmon are acceptable for sale in Canada or export.

In order for the study to be considered a success, the canned product must include a range of product which would be rated from accept to reject. In addition, a majority of the product must fall between these two extremes, in the "grey" or "borderline" zone. This is important because it is in the "grey zone", between accept and reject, where much discussion and examples of cans will be required in order to establish a firm "cut off" point.

Both Inspection branch of DFO and members of the BC fishing industry contributed resources for the study. A contribution by the Province of British Columbia (BC) to the Industry participants is also acknowledged. DFO Inspection provided staff to set up the spoilage run, oversaw the landing and evaluation of the fish, and participated where necessary in the processing of the fish. Graders were provided to assess the fresh fish and the final canned product, and several individuals were available as required, to assist with tasks including handling and butchering the fish. Chemical evaluations of the fresh and canned product were also conducted. Ethanol concentration, which reflects the degree of bacterial decomposition, has been shown in earlier studies to correlate with the sensory evaluation of salmon (Hollingworth and Throm 1982, 1983 Hollingworth et al 1987). Cadaverine and putrescine have been shown to increase in concentration during the decomposition of salmon (Mietz and Karmas 1978, Karmas 1981, Yamanaka et al 1989). Total volatile base nitrogen (TVB-N) has been used to evaluate the quality of various seafoods (Malle et al 1983 and Lehmann 1997), however a number of factors limit its use (Rehbein and Oehlenschlaeger 1990).

Industry provided the fish for the study, the facilities and personnel to process the fish, and the cans.

The overall action plan for completion of the study was divided into five sections:

Landing and Evaluation of Fresh Fish;
Spoilage of Fish, Grading and Sampling;
Processing;
Chemical Analysis of Fresh and Canned Product;
Sensory Assessment of Final Canned Product.

2.0 Landing and Evaluation of Fresh Fish

The study was divided into two time periods. The first, in late July 1994 in Prince Rupert, was selected because pink and sockeye salmon were being fished at the same time. Canadian Fishing Company (CFC) held the fish for the duration of the spoilage study, and BC Packers Ltd. canned the fish. The second half of the study took place in November 1994, in Vancouver and used chum salmon from the Fraser and Qualicum Rivers. In the original design of the experiment, both silver bright and late chum were to be collected and spoiled. This plan had to be changed when it was found that there were insufficient silver bright chums available at the time of the study. Instead, late chum from the Fraser and Qualicum Rivers were compared. One hundred pounds of frozen silver bright chum which lacked the distinct "late" odour and flavour which persists in the more mature fish was also canned for use by the sensory panellists and for chemical analysis. Ocean Fisheries Ltd. provided the fish, the space to hold the fish during the spoilage study, the canning facilities and the staff and labour to process the fish.

After completing the spoilage study in Prince Rupert, the procedures used for holding and handling the fish were reviewed. In situations where problems were identified, procedures were changed and improved. Therefore, in some cases, the pink and sockeye salmon were handled in a slightly different manner than the chum.

2.1 Initial Grading and Sampling of the Product

The fish selected for the study were chosen according to strict criteria. The best possible quality Canadian salmon were selected from vessels with well-iced fish holds or with chilled seawater (CSW) systems in place. As the fish were unloaded from the fishing vessel, company graders first sorted them. The fish were then re-evaluated by a DFO grader and rejected for seal bites, belly burn, or exterior flaws, which had damaged the integrity of the skin. Sockeye and pink salmon showing signs of sexual maturity were also rejected. Using these criteria, approximately two out of three sockeye and pink were rejected. It was difficult to maintain a constant level of maturity in the pink salmon; some of the boats had distinctly more mature fish than others. As a result, some of the pinks selected were slightly more mature than others. A random selection of male and female fish was used. Chums of the same approximate maturity were selected using the criterion of similar exterior appearance. This included hooked noses, and distinct colour bars along the sides of the fish. Approximately one in three chum salmon were rejected during the initial grading. Exterior flaws such as cuts, and extensive bruising were less evident in the chum, and the level of maturity initially appeared to be quite consistent. During later grading however, different maturity levels were apparent by the distinct differences in flesh colour. Some of the fish had firm, pink flesh while in others; the flesh had become white and soft. Only male fish were provided for both the Qualicum and Fraser River chums.

The pink and sockeye salmon were generally held on the packer vessel for one or two days before being delivered to the loading station. Much of the pink and chum salmon were in good condition, for example, some were still in rigor during unloading. The sockeye were in poorer condition than the other two species, all of the fish had completed rigor, and some had not been held at an optimum temperature.

3.0 Spoilage of Fish, Grading and Sampling

3.1 Spoilage procedures

After initial grading, the fish were weighed and then transferred into plastic insulated totes, which were fitted with lids and drainage. The fish intended for the iced study were held in a slush of water and layered ice. Throughout the "iced" study, the totes were partially drained to allow access to the fish, and were then top iced. Pink and sockeye used in the ambient study were held in water in an attempt to both provide buoyancy to prevent crushing of the fish on the bottom of the tote, and to maintain a consistent temperature throughout the tote. It was observed however that the water caused the fish to putrefy in an "uncharacteristic" manner. To avoid this "uncharacteristic putrefaction", in the fall study, the ambient chum were not held in water for the entire study. Upon arrival at the plant, the ambient batch of chum was held in water overnight to equilibrate the temperature throughout the tote as quickly as possible. The water was then drained and the fish were held "dry" for the duration of the study. It is difficult to compare these two methods because the ambient temperature in the summer, 14 to 15C was distinctly different from the fall temperature of approximately 7C. Table 1 outlines the amount of fish used overall, for each study, and for each increment.

The amount of fish drawn at each increment was intended to produce four cases of each species. The final cans produced are intended for use by several DFO/industry workshops, by the Industry themselves for internal training needs and standardisation, and for chemical analysis. In addition, an inventory of cans will be retained as a library for future standardisation.

A final tally showed that a sufficient number of cans were produced for most of the codes. Only two cases of PA120 and SA120 remain because all of the cans had peaked during processing, and most were later destroyed. The Qualicum and Fraser River chum increments generally have at least two cases each because the initial amount of fish was divided between the two runs.

3.2 Temperature Monitoring

The temperature of the fish at the time of unloading varied somewhat, but was generally between 1 to 3C. Stainless steel thermocouple probes were used to measure the internal temperature of the fish during each experiment. For the pink and sockeye, the probe was inserted under the skin along the lateral line of the fish. For the Fraser chum, the probe was inserted into the belly cavity as this was considered to provide better information regarding the condition of the fish. The Qualicum chums were not monitored for temperature changes. Thermocouple wires were also placed throughout several totes to monitor for temperature gradients.

3.3 Results of the temperature monitoring

In Figure 1, the temperature of a tote full of Fraser River chum salmon at ambient temperature and the temperature fluctuations (late October-mid November) in the same room are compared. Notice the gradual increase in the temperature within the tote from 4-5C to 9C, despite the widely fluctuating temperatures in the room (2-10C). In Figure 2, the abdominal temperatures of chum salmon held in ice and at ambient temperature are compared. Notice that while the iced fish maintain a steady temperature of approx. -1C, there is a gradual increase in the abdominal temperature of the ambient fish from 5C to 9C. In Figure 3, the average change in abdominal temperature of sockeye and pink are shown during ambient and iced storage. At ambient temperature, apart from the fact that the pink salmon were initially at a lower temperature (5C) than the sockeye (14C), the temperature remained stable between 14-15C. This may have been a result of holding the fish in water for the entire study. The greater fluctuations in the iced tote reflect thermogenic metabolic activity within the digestive systems of these fish, which had been actively feeding at the time of capture.

3.4 Sampling during spoilage

Sampling occurred at the times outlined in Table 2. Because the objective of the study was to produce cans of salmon with an incremental increase in decomposition, the sampling periods were constantly adjusted based on the perceived rate of spoilage, and the temperature at which the fish were held. Table 2 shows that the four groups of fish were not always sampled at the same times. For example, several increments of the ambient sockeye were not collected because of scheduling difficulties. The chum salmon were held at ambient temperature much longer than the sockeye and pink salmon because of temperature differences. At 14 to 15C, the fish were held for 120 hours i.e. 1680-1800 degree hours, while at 6 to 7C, the fish were held for 264 hours i.e.1584-1848 degree hours. The Qualicum chum were held two days longer than the Fraser River chum, again this was a function of temperature. The ambient temperature was several degrees colder at the beginning of the Qualicum spoilage study than the Fraser River study. The chum were also sampled at longer intervals than the sockeye or pink in order to reach a final "end point" where all of the canned fish would exhibit characteristics of advanced spoilage.

3.5 Grading

Twenty-one pink and sockeye, and six chum salmon were graded during each increment. The small sample size for the chum was a result of the physical size of the fish, which were 6.4 Kg on average, and because the study was divided in half to accommodate the two runs. After drawing the fish from the totes, they were identified by a piece of labelled surveyor's tape tied around their tails. During the pink and sockeye study, one grader performed all of the sensory analysis on the fresh fish up until the end of the ambient study when a second grader took over. During the chum study, two graders worked together throughout the entire study. Results of the grading were reported using a 10 cm linear scale which ranged from acceptable (0 to 5) to reject (>5 to 10), and written descriptions.

3.6 Results of the fresh fish grading

Pink salmon: The iced pink salmon at 0 hour were all acceptable, and were marked at approximately 0.5 on the linear scale. On day 12, 5% of the fish (1 out of 21) were rejected. By day 14, only 9.5% of the fish were acceptable, the rest were rejected. Typical descriptions used to describe the fish were belly burn, ribs protruding, musty, sour odour, and soft texture.

The ambient pink salmon at 0 hour were quite variable. Two fish were rejected for ammonia odour, and the acceptable fish ranged from 0.8 to 4.8 on the linear scale. By 72 hours, only 19% of the fish were accepted, and by 96 hours, all of the fish were rejected. Typical descriptions of the fish were sour, sulphide, and ammonia odour, and belly burn.

Sockeye salmon: The iced sockeye at 0 hour were all acceptable, ranging from 0.3 to 0.45 on the linear scale. By day 15, 15% of the fish were rejected, and 85% were borderline accept. The fish were found to have severe belly burn, and protruding ribs.

The ambient sockeye at 0 hour were all accepted, and ranged from 0.5 to 0.7 on the linear scale. By 24 hours, one fish was rejected, and by 72 hours, all 21 fish had been rejected. Typical descriptions of the fish included belly burn, sour, ammonia and sulphide odour.

Fraser River chum salmon: The day 0 iced Fraser River chum were all acceptable with a "characteristic" odour. On day 19, all fish were acceptable except one, which was on the borderline. The fish were described as stale, late, and "turnipy". On day 21, all of the fish were rejected. The study continued to day 26; by this time the fish were marked between 8 to 9 on the linear scale, and were described as having sulphide, strong turnip, and sour milk odours. The belly cavity was flushed in some cases, but no belly burn was noted.

The 0 hour ambient Fraser River chum were acceptable, and were described either as having a characteristic odour, or as slightly stale. By 120 hour, 90% of the fish were rejected. The fish were described as musty, sour, stale, neutral and late. By hour 144, all of the fish were rejected, and had hydrogen sulphide, putrid, or hydrocarbon odours. Hour 168 was the last increment for this study, the fish were described as having acrid, putrid, and hydrocarbon type odours.

Qualicum River chum salmon: The day 0 iced Qualicum chum salmon were acceptable, and described as late with dark skin, no belly burn, and a firm texture. On day 21, 20% of the fish were rejected, and on day 23, all of the fish were rejected. This contrasts with the chemistry data, which indicates that, the iced Qualicum chum were more advanced in spoilage than the Fraser chum from day 19 onwards. On day 28, the last day of the study, the fish were described as having hydrocarbon, hydrogen sulphide, sour, "turnipy" odours, and severely flushed belly walls.

The 0-hour ambient chum salmon were all acceptable, but described as slightly late. By 192 hours, all of the fish were rejected, and had odours including hydrogen sulphide, sour, late, "bilgy", and putrid.

4.0 Processing

The fish were dressed and cannery butchered by a small crew in an area separate from other processing. It was important to keep the fish used in the study separate from other production so that their identity could be maintained, and to prevent the fish used in the latter part of the study from contaminating the equipment in the facility. A list of the codes used throughout the study is presented in Table 3, and includes the identification key used to label the can ends. After being closed, the cans were retorted, held overnight in the post process area, and then boxed or palletised. Note that many of the ends were precoded, and in the event, processing times were delayed, the time frames noted may not be completely accurate (Table 3).

5.0 Chemical Analysis of the Frozen and Canned Product

5.1 Sampling for Chemical Analysis

On day 0 of each spoilage study, five salmon (sockeye, pink or chum) were dressed and frozen for use as a reference for chemical analysis. During each increment, 10 steaks of each species were frozen and held in cold storage before being shipped to Halifax for total volatile base nitrogen (TVB-N) analysis. The day 28 Qualicum chum samples were lost due to equipment breakdown. Five cans each of pink and sockeye were analysed for ethanol, cadaverine and putrescine at each of the shorter increments and ten cans of each at the longer storage increments. Five cans were sampled for each increment of the two chum runs.

5.2 Procedures for Chemical Analysis

After shipping to Halifax the frozen salmon steaks were analysed for total volatile base nitrogen (TVB-N) using an MgSO₄ extraction method, adapted to an automated, Tecator 1043 Autoanalyser (Woyewoda et al 1986). Canned salmon at each spoilage increment was analysed for ethanol, cadaverine and putrescine. Ethanol was quantified using an automated, headspace sampling, gas chromatographic method with mass spectrometer detection of the two major ions of ethanol (McLachlan et al 1998). The biogenic amines, cadaverine and putrescine were analysed according to the DFO Chemical Methods Manual (DFO Inspection 1986).

5.3 Results of Chemical Analysis

Sockeye salmon

The concentrations of TVB-N in sockeye steaks that had been taken from fish held at ambient temperature are shown in Figure 4. The TVB-N concentration rose from 10mg/100g at zero time to >20mg at 72 hrs and >35mg at 120 hrs. The concentrations of TVB-N in sockeye steaks taken from fish held in slush ice are shown in Figure 5. The mean TVB-N concentration in these fish fluctuated between 10 and 17mg/100g and only a few individual samples exceeded 20mg. Thus, we can conclude that while TVB-N may be a useful indicator of deterioration in sockeye held at ambient temperature it is not a satisfactory indicator of deterioration in sockeye held on ice.

Concentrations of ethanol and the two diamines, cadaverine and putrescine in sockeye held at ambient temperature and in slush ice before canning are shown in Figure 6. Clearly all three parameters rose more quickly in the sockeye held at ambient temperature compared to those held on ice.

Ethanol rose from <40 ppm after 24 hrs at ambient temperature to 200ppm at 48 hrs and >350 ppm at 120 hrs. Similarly cadaverine increased from <15ppm after 24 hrs at ambient temperature to >70 ppm at 48 hrs and >180 ppm at 120 hrs. Putrescine concentrations rose more slowly from <5 ppm at 24 hrs to 10 ppm at 48 hrs and 40 ppm at 120 hrs.

In slush ice, the ethanol concentration did not reach 50 ppm until day 13 of storage and on day 19 had reached 90 ppm. Cadaverine had risen to 20 ppm by day 11 in ice and rose to >70 ppm by day19. Putrescine rose very slowly in sockeye held in ice reaching <20 ppm at day 19. We can conclude that both ethanol and cadaverine are good indicators of deterioration in sockeye with putrescine being a less useful indicator especially in iced sockeye.

Pink salmon

The concentrations of TVB-N in pink salmon steaks that had been taken from fish held at ambient temperature are shown in Figure 7. The TVB-N concentration rose from 15mg/100g at zero time to >20mg at 72 hrs. and >35mg at 120 hrs. The concentrations of TVB-N in sockeye steaks taken from fish held in slush ice are shown in Figure 8. The mean TVB-N in these fish varied between 7 and 10 mg/100g at zero to 12 days and then rose gradually to 20mg at 20 days. The 14-day sample demonstrated greater TVB-N variability than the other sample intervals. Thus in pink salmon, as in sockeye, TVB-N may be an indicator of spoilage in ambient held fish but not in salmon held on ice.

Concentrations of ethanol, cadaverine and putrescine in pink salmon held at ambient temperature and in slush ice, prior to canning, are shown in Figure 9. In the pink salmon, as in the sockeye, all three parameters increased more rapidly and to higher levels in the ambient temperature set than in the iced set. Thus in pink salmon held at ambient temperature ethanol rose from >30 ppm at 24 hrs to >60 ppm at 60 hrs and >370 ppm at 120 hrs. In these fish cadaverine had risen to 20 ppm at 24 hrs to >45 ppm at 60 hrs and to 200 ppm at 120 hrs. Putrescine on the other hand rose gradually from <5 ppm at 0 hrs to <20 ppm at 96 hrs and then rose rapidly to >40 ppm at 120 hrs.

When pink salmon were held in slush ice the ethanol concentration in the canned product did not rise above 20 ppm until after day 12 while on days 18 and 20 the ethanol concentration was 100-120 ppm. The cadaverine concentration in pink salmon held in ice rose slowly from very low levels on days 0 and 3 to >10 ppm on day 12. It then increased rapidly to 50-80 ppm on days 14-20. The putrescine concentration in these fish rose very gradually from <5ppm on days 0 and 3 to >10 on day 20. We can conclude that both ethanol and cadaverine are good indicators of spoilage in pink salmon with putrescine being a less useful indicator especially during iced storage.

Chum salmon

Concentrations of ethanol, cadaverine and putrescine in Qualicum chum salmon held at ambient temperature and in slush ice, prior to canning, are shown in Figure 10. In the chum salmon changes in the chemical parameters were generally more gradual than in sockeye and pink salmon despite the longer storage periods. The exception being the last sampling of the ambient set at 264 hrs. These gradual changes were probably associated with the lower ambient temperature at the time of the chum study. In the chum salmon held at ambient temperature ethanol rose gradually from 10 ppm at 0 and 48 hrs to >40 ppm at 192 hrs and then rapidly to >160 ppm at 264 hrs. Cadaverine in chum salmon held at ambient temperature remained below 5 ppm until 96 hrs and then rose to 30ppm at 192 hrs and then rapidly to >210 at 264 hrs. Putrescine in these same samples was very low until 96 hrs and was still below 10 ppm at 192 hrs. However by 264 hrs it had risen to >25 ppm.

In pink salmon held in slush ice ethanol remained below 20 ppm until day 19 and rose to 50 ppm by days 23-28. In these fish cadaverine was <5 ppm until day 14 and gradually rose to >20 ppm by day 23. Putrescine on the other hand was initially very low rising to 5 ppm at day 14 and to almost 10 ppm at day 23. The quantitative increases in ethanol in chum salmon during storage were less than those seen in pink and sockeye salmon.

6.0 Sensory Assessment of the Final Canned Product

6.1 November 1994 DFO/Industry Canned Salmon Grading Orientation Session

In November 1994, a Canned Salmon Grading Orientation Session was held. Participants included eight members of the Industry, three DFO Inspection personnel and four observers. The primary objective of this one-day session was to familiarise Industry personnel with DFO’s method of evaluating fish products using a Taint Decomposition Unwholesome (TDU) standard in preparation for participation in DFO/Industry workshops to determine accept/reject levels in canned salmon. Procedures used in this orientation would become the standard procedures used in subsequent DFO (CFIA)/Industry canned salmon sensory workshops.

Participants were provided with a series of presentations consisting of DFO sensory workshop training methods, the evaluation of sensory abilities against chemical testing parameters, development of attribute vocabulary and instruction on the use of an unstructured 10-centimeter line scale. A draft version of the Canned Salmon Standard (DFO Inspection Directorate 1996), prepared by the Canned Salmon Standard Working Group (consisting of Industry and DFO members), was also provided as a reference. Each analyst was assigned a personal identification number to be used on the ballots. This number was subsequently used in any discussion of the data during the workshop. A schedule was also provided which indicated not only the approximate start time of each sensory set, but the type of sensory set (discussion or blind) and the species involved.

Prior to entering the sensory lab where the samples were presented, analysts washed their hands and were given a ballot for the sensory set they were about to evaluate. Ballots identified the sensory set number, type of sensory set (discussion or blind) and the species involved.

Three sets of fifteen cans each, comprised of the quality range produced from each species and commercial samples, were presented to the participants for sensory evaluation. Samples were identified by three-digit random numbers. Ballots required the analyst to indicated whether the sample was acceptable or unacceptable (pass/fail) and to provide a sensory rating on an unstructured 10-centimeter line scale. The ballots also had a field for comments where analysts were encouraged to provide descriptors and other notes.

The unstructured 10-centimeter line scale is a tool, which was used by DFO Fish Inspection and is now used by the Canadian Food Inspection Agency in sensory workshop/training sessions. The line has two endpoints and a mid-point. The analyst will draw a line a vertically across the line scale to the left of the mid-point if he/she believes the sample is acceptable (0 cm - 4.9 cm). If the analyst believes the sample is unacceptable, a vertical line is drawn to the right of the mid-point (5.1 cm - 10 cm). The better the quality of the sample, the farther to the left is the vertical line drawn. The poorer the quality of the sample, the farther to the right is the vertical line drawn. The analyst may use the endpoints but the mid-point is out of bounds (i.e. the analyst must make a decision to pass or fail the sample). The closer to borderline a sample is with respect to quality acceptance, the closer the line to the mid-point. In this manner, the 10-centimeter line scale allows an analyst’s sensory evaluations to be quantified. The line score values can then be graphed to provide a visual interpretation for training and calibration purposes.

Sensory booths were utilised to provide analysts privacy and minimise distractions. Each booth contained one sample and analysts walked from booth to booth until all fifteen samples had been evaluated. Revisiting a sample is permitted and the sequence in which analysts evaluate the samples were randomised since analysts were instructed to move to any vacant booth they have not visited yet (or wished to revisit). Water for rinsing and plain unsalted soda crackers for cleansing the palate between samples were made available.

Once an analyst had completed the evaluation of all fifteen samples, the ballot was returned to the workshop co-ordinator or assistant. The analyst then moved out of the sensory lab to a boardroom to await the next sensory set. Analysts were instructed not to discuss their results for blind sensory sets between one another.

6.2 Results of the Sensory Assessment

As this session was meant to be only an orientation, extensive data evaluation of the sensory results of individual analysts was not performed beyond general comparisons of agreement between Industry and DFO. There was strong Industry agreement with DFO in the sockeye sensory set although as samples approached the accept/reject cut-off point (borderline samples), the percentage of agreement dropped (Table 4). When the mean scores for the industry analysts were plotted against mean scores for DFO for the sockeye samples it can be seen that there was good overall agreement (Figure 11). However, when an individual industry analyst was plotted against the DFO analysts it can be seen that this industry analyst was more cautious in accepting samples than the DFO analysts (Figure 12). The strongest Industry agreement with DFO occurred in the pink salmon sensory set (Table 5). The chum salmon sensory set showed the lowest percentage agreement between Industry and DFO with five of the fifteen samples with less than 60% agreement (Table 6). Fraser chum salmon on ice at day 0 (KI00) and Qualicum chum on ice at day 3 (KXI03) are included in this group of five.

6.3 January 1995, Internal DFO/ Canned Salmon Workshop

In preparation for the January 1995 DFO/Industry Canned Salmon Grading Workshop, an internal workshop was held to establish DFO’s position on the accept/reject levels in canned salmon using a Taint Decomposition Unwholesome (TDU) standard. The internal workshop also served to standardise the DFO graders. Participants consisted of eight DFO inspectors. Ten sets of fifteen cans each (three sets each of sockeye and pink salmon and two sets each of Fraser chum and Qualicum chum salmon) were presented to the participants for sensory evaluation. Samples were comprised of the quality range produced from each species and commercial samples. Procedures for the workshop followed those established for DFO sensory workshops/training sessions and are outlined in the section describing the November 1994 DFO/Industry Canned Salmon Grading Orientation Session.

The accept and reject scores for the eight analysts are shown in Table 7 for sockeye, in Table 8 for pink, and in Table 9 for chum salmon. Several analysts had numerous false accepts in the sockeye and pink sensory sets which lowered their percentage score on reject units and subsequently their overall percentage correct. The opposite tendency was observed for the Fraser and Qualicum chum salmon sensory sets where several analysts had numerous false rejects, which lowered their percentage score on accept units and subsequently their overall percentage correct.

For the purpose of this exercise, the "Expert decision" was the history of the spoilage sample and the ethanol concentration. This helped identify the spoilage treatment groups that were approaching the accept/reject cut-off level with respect to a Taint Decomposition Unwholesome (TDU) standard and also spoilage treatment groups with a greater quality variation. Sockeye at ambient for 48 hours (SA48) had an ethanol level of 155.5 ppm., but in three of five replicates, less than half of the analysts rejected the sample. Sockeye on ice at 13, 17 and 19 days had ethanol levels of 49.7 ppm., 71.9 ppm. and 93.5 ppm. Respectively and also received mixed sensory evaluations. Pink salmon at ambient for 60 hours had an ethanol level of 43.8 ppm. A majority of the analysts accepted the sample in four of the five replicates. Chum salmon at ambient for 120 hours from both the Fraser and Qualicum runs were usually rejected by a large majority of analysts but the ethanol concentrations in these salmon were only 18.26 ppm and 17.29 ppm respectively. It can therefore be concluded that the cut off level for ethanol in sockeye and chum is in the region of 50 ppm while the cut off level for ethanol in chum salmon is in the region of 17 ppm (Schenkeveld et al 1999).

6.4 January 1995, DFO/Industry Canned Salmon Workshop

A DFO/Industry Canned Salmon Grading Workshop was convened in January 1995. Participants included ten members of the Industry, three DFO Inspection personnel and four observers. The primary objective of this two-day session was to determine accept/reject levels in canned salmon using a Taint Decomposition Unwholesome (TDU) standard. The workshop began with an overview of the spoilage work and the resulting spoilage samples. Highlights of sensory results and chemical analysis of the spoilage samples by Inspection staff were provided. A total of eight discussion sets, one of each species and treatment group, were presented to analysts in sequential order. The purpose of these discussion sets was to present to the analysts the full sensory quality range of the spoilage samples to be used for the workshop. The sets also provided analysts with an opportunity to calibrate their use of the 10 cm line scale being used on the ballots. Following the discussion sets six blind sensory sets were presented to analysts for evaluation. Each blind set consisted of fifteen cans, comprised of the quality range produced from a given species and commercial samples. A three digit random number identified samples for these blind sets.

Two sets of sensory analysis data are presented from this workshop. First, the line score values for the DFO analysts under non-blind conditions are shown for each species at each spoilage increment during ambient and iced storage. The line score values for sockeye at ambient temperature show that any cans prepared from sockeye held more than 48 hours was not acceptable while sockeye held on ice was acceptable up to 8 days, was variable in its assessment at 11 and 13 days and was judged by almost all inspectors to be unacceptable after 13 days (Figure 13). The line score values for pink salmon at ambient temperature show that cans prepared from pinks held more than 36 hours was marginal or not acceptable while pink held on ice was acceptable up to 12 days, was variable in its assessment at 14 and 16 days and was judged by all inspectors to be unacceptable after 16 days (Figure 14). The line score values for Fraser chum salmon at ambient temperature show that cans prepared from chum held more than 72 hours was marginal or not acceptable while chum held on ice was marginally acceptable up to 19 days, and was judged by all inspectors to be unacceptable after 19 days (Figure 15). The line score values for Qualicum chum salmon at ambient temperature show that cans prepared from Qualicum chum, even at zero time, were only marginally acceptable to 3 out of 4 analysts. Chum held more than 120 hours was not acceptable. Qualicum chum held on ice was marginally acceptable up to 14 days, and was judged by all inspectors to be unacceptable after 21 days (Figure 16).

The second data set presented from this workshop is the line score range for each sample unit for four DFO analysts and ten industry analysts. Overall, the data indicates good agreement between the DFO and industry analysts. The line score values for sockeye at ambient temperature show that cans prepared from sockeye held up 48 hours were acceptable to almost all analysts. Sockeye held on ice was acceptable up to 5-8 days, was variable in its assessment at 11 and 13 days and was judged by almost all inspectors to be unacceptable after 13 days (Figure 17). The line score values for pink salmon at ambient temperature show that cans prepared from pinks held more than 36 hours was marginal or not acceptable. Pink salmon held on ice was acceptable up to 12 days, was variable in its assessment at 14 and 16 days and was judged by all inspectors to be unacceptable after 16 days (Figure 18). The line score values for Fraser chum salmon at ambient temperature show that cans prepared from chum held up to 21 hours was acceptable to all analysts. Chum held 72-120 hours was marginal or not acceptable. Fraser chum held on ice was, depending on the analyst, acceptable to marginally acceptable up to 19 days, and was judged by all inspectors to be unacceptable after 21 days (Figure 19). The line score values for Qualicum chum salmon at ambient temperature show that cans prepared from Qualicum chum, even at zero time, were only marginally acceptable to several analysts and chum held more than 120 hours was not acceptable. Qualicum chum held on ice was marginally acceptable to most analysts up to 14 days, and was judged by all inspectors to be unacceptable after 23 days (Figure 20).

The rapid deterioration in the acceptability of sockeye and pink salmon held at ambient temperature reflects the higher ambient temperature that these fish were held at compared with the chum salmon that were studied in the fall when the ambient temperature was lower. From a sensory standpoint, the ambient temperature sockeye underwent a clear transition from acceptable to unacceptable whereas the iced sockeye appeared to pass through a stage at 11-13 days of storage which evoked a mixed response from both DFO and industry analysts. Ambient held pink salmon also went through a clear sensory transition. Iced pink salmon underwent a consistent progressive transition from acceptable to not acceptable remaining acceptable slightly longer than the sockeye. Chum salmon, especially the Qualicum chum, was only marginally acceptable to some analysts even at the beginning of the ambient or iced storage periods. This may have been a result of the chum having characteristic late odour that resulted from sexual maturation (Hatano et al 1986).

7.0 Conclusions

Spoilage sets of canned sockeye, pink and chum salmon have been prepared which reflect the effects of controlled deterioration in both ambient temperature and iced storage. The changes in four chemical parameters, ethanol, cadaverine, putrescine and TVB-N have been quantified allowing comparison with the sensory evaluation. Ethanol, cadaverine and to a lesser extent putrescine displayed clear-cut increases during ambient and iced storage. TVB-N proved to be a less reliable indicator of spoilage especially during iced storage. Comparison of the results of sensory evaluation of these spoilage sets by DFO and industry analysts revealed a good concurrence of line score accept reject evaluation.

8.0 References

DFO Inspection, 1986. Chemical Methods Manual.

DFO Inspection Directorate, 1996. Fish Products Standards and Methods Chapter 2, Standard 8, Canned Salmon Standard 7pp.

Griffin, N. 1998. The sweet smell of seafood. Seafood International January 1998 pp. 29-30.

Hatano, M., K. Takahashi, and K. Zama. 1986. Quality evaluation as food of chum salmon captured during spawning migration. Nippon Suisan Gakkaishi 53: 479-486.

Hollingworth, T.A.,Jr. Throm, H.R 1982. Correlation of ethanol concentration with sensory classification of decomposition in canned salmon. Journal of Food Science. 47: 1315-1317.

Hollingworth, T.A.,Jr. Throm, H.R 1983. A headspace method for the rapid analysis of ethanol in canned salmon. J. Food Sci. 48: 290-291.

Hollingworth, T.A.,Jr. Throm, H.R. Wekell, M.M 1987. Determination of ethanol in canned salmon. Developments in food science. Amsterdam. no. 15. Available as: SEAFOOD QUALITY DETERMINATION. pp. 153-160.

Karmas, E. 1981. Biogenic Amines as Indicators of Seafood Freshness. Lebensmittel Wissenschaft Technologie. 14: no. 5, pp. 273-275.

Lehmann, I. 1997. Pre-packaged fresh fish -- searching for quality descriptive criteria. Informationen fur die Fischwirtschaft. Hamburg.. 44, no. 4, pp. 182-185.

Malle, P. Nowogrocki, G. Tailliez, R 1983. Relationship between deterioration index and rates of total volatile basic nitrogen for marine fish. [French] Sciences des Aliments. 3, no. 2, pp. 311-322.

McLachlan, D. G., P. D. Wheeler and G. G. Sims, 1998. An automated gas chromatographic method for the determination of ethanol in canned salmon. J. Agric. Food Chem. 47: 217-228.

Mietz, J.L. and E. Karmas 1978. Polyamine and histamine content of rockfish, salmon, lobster, and shrimp as an indicator of decomposition. JAOAC 61:139-145.

Rehbein, H. Oehlenschlaeger, J. 1990. (On the determination of TVBN: Chemical basis and limitation of the method.). [German] LEBENSMITTELCHEM. GERICHTL. CHEM. 44: 26-29.

Rielly, T.I. York, R.K. 1994. Sensory analysis application to harmonize expert assessors of fish products. In Sylva, G. Shriver, A.L. Morrissey, M. Eds. Quality control and quality assurance for seafood, 16-18 May 1993, Newport, OR (USA). Oregon Sea Grant Publ. p. 120-124.

Schenkeveld, S., D.G. McLachlan, R. Reid, C. Cosham, P. Wheeler, P. Chan, and G.G. Sims. 1999. Correlation of sensory attributes with ethanol content in canned Pacific salmon. In preparation.

Woyewoda, A. D., S. J. Shaw, P. J. Ke and B. G. Burns. 1986. Recommended laboratory methods for assessment of fish quality. B. Quality Indices- Refrigerated and iced storage. 10. Total Volatile Base. Canadian Technical Report of Fisheries and Aquatic Sciences # 1448, pg. 46-49.

Yamanaka, H. Shiomi, K. Kikuchi, T 1989. Cadaverine as a potential index for decomposition of salmonoid fishes. [Japanese] Journal of the Food Hygienic Society of Japan. 30, no. 2, pp. 170-174.

York R.K. 1989. Canadian fish products-fish inspection and sensory evaluation. Can. Inst. Food Sci. Technol. J. 22: 441-444.

Table 1. Weight (Kg) of pink, sockeye, and chum salmon used in each of the studies.

Table 2. Sampling schedule for the canning of the three species of salmon held under ambient and iced storage conditions.

* The 0-hour sampling period occurred approximately 36 hours after the fish were landed.
** The 0-hour sampling period occurred approximately 24 hours after the fish were landed.
1 The 0-hour sampling period occurred approximately 12 hours after the fish were landed.
11 The 0-hour sampling period occurred approximately 4 hours after the fish were landed.

Table 3. List of can codes used for the three species.

Identification Key

P    pink salmon (Oncorhynchus gorbuscha)
S    sockeye salmon (Oncorhynchus nerka)
K    keta (chum salmon) (Oncorhynchus keta) from the Fraser River
KX keta (chum salmon) (Oncorhynchus keta) from the Qualicum River
KS silver bright (chum salmon) (Oncorhynchus keta) previously frozen
I      iced storage
A    ambient storage
       Numerals represent time of holding in hours for ambient and in days for iced.

* In some instances because of restrictions to processing times, the code information may not be completely accurate. Thus, the PA36 and PA60 codes were processed at approximately 30 hours and 52-54 hours respectively

Table 4. Results of the Sockeye Salmon Sensory Set at the November 1994 Orientation Workshop.

Table 5. Results of the Pink Salmon Sensory Set at the November 1994 Orientation Workshop.

Table 6. Results of the Chum (keta) Salmon Sensory Set at the November 1994 Orientation Workshop (K= Fraser River chum, KX= Qualicum River chum).

Table 7 . Accept and reject scores for sensory evaluation of canned sockeye salmon spoilage sets by eight DFO Inspectors during Internal DFO canned salmon workshop January 1995.

Table 8 . Accept and reject scores for sensory evaluation of canned pink salmon spoilage sets by eight DFO Inspectors during Internal DFO canned salmon workshop January 1995.

Table 9. Accept and reject scores for sensory evaluation of canned chum salmon spoilage sets by eight DFO Inspectors during Internal DFO canned salmon workshop January 1995

Figure 1 - click graphic for larger image

Figure 2 - click graphic for larger image

Figure 3 - click graphic for larger image

Figures 4 to 12 (click on link to download Excel file)

Figures 13 to 20 (click on link to download Excel file)

Top of Page

Important Notices