Pink salmon are usually too small to market as steaks and the texture of the flesh is softer than that of other species of salmon, which makes them more difficult to fillet. Use of mechanical deboning may provide pink salmon flesh that could be processed into marketable, convenient forms on an economical basis. Accordingly, the objectives of this experiment were to determine the yield at various steps in the processing of pink salmon into boneless fillets or mince, to determine changes in quality during frozen storage of product forms made from fillets and/or mince, and to determine if frozen, headed and gutted, pink salmon can be reprocessed into these forms.
Pink salmon were caught by a commercial seiner in the vicinity of Kodiak Island on 28 July 1986 and transported in slush ice to a processor in the city of Kodiak, Alaska. The average weight of the fish was 1.6 kg. The next day, about 1,000 kg of iced salmon were transferred to the National Marine Fisheries Service (NMFS) Utilization Research (UR) Division pilot plant facilities in Kodiak where heads and internal organs were removed by hand. Collars were left on the fish. Most of the headed and gutted (H&G) fish were frozen either in a plate freezer for 6-12 hours at -40°C or in an air blast freezer for 16 hours at 0°C, glazed in fresh water, packaged in 2.0-mm polyethylene sleeves, and placed in a master carton lined with a polyethylene bag. The remaining unfrozen H&G fish were re-iced.
The next day, the unfrozen H&G fish were trimmed to remove the backbone, and then fins, collars, and rib bones were removed. The fish were mechanically skinned using a Baader Model 50 skinner [use of trade names does not imply endorsement by the National Marine Fisheries Service, NOAA]. Pin bones were removed by making parallel cuts the length of each fillet (Fig. 1). Boneless, skinless fillets were placed in polyethylene bags and iced overnight. Trimmings from removal of pin and rib bones were similarly treated. The next day, trimmings and some of the fillets were minced using a Baader Model 694 deboner equipped with a 5-mm drum. Blocks containing 0 (fillets), 25, 50, or 100% mince were prepared using the method of Babbitt et al. (1987) and frozen at -40°C in the plate freezer. The frozen blocks were placed in master cartons and stored at -18°C as were the H&G fish. Several of the fillet blocks were stored at -34°C to serve as a control sample for the sensory evaluations. Yield data were collected at each stage of the sample preparation. Yields at various steps in commercial processing of salmon were obtained at the processor's plant on one lot of 10 fish. In the commercial operation, the viscera were removed manually, heads and collars were removed by a Baader Model 421, filleting was done with a Baader Model 184, skins were removed with a Baader Model 50, and trimming was done by hand. These blocks, made from unfrozen H&G fish, are referred to as once-frozen blocks in the discussion of results.
At 3, 6, and 12 months of frozen storage, H&G fish were thawed overnight at room temperature (60-65°C) prior to being manually filleted, skinned, and trimmed. The fillets and minced flesh were immediately processed into the various block forms previously described and refrozen. These twice-frozen blocks are referred to as reprocessed blocks in the discussion of results.
Thus, our samples consisted of H&G pink salmon that were made into fillet blocks (0% mince) and blocks containing 25, 50, and 100% mince. The mince was made from pink salmon fillets and trimmings of pin and rib bones. These blocks were then frozen. One test group used fresh H&G fish known as once-frozen blocks. The second test group used frozen H&G fish that had been in storage for 3, 6, or 12 months. These twice-frozen blocks are the reprocessed blocks.
To determine thaw drip, an unthawed portion of the same size as that used for sensory analysis was weighed and placed in a perforated plastic bag which was placed inside an outer bag weighted to insure submersion. The samples were immersed in a 15°C bath for 40 min, and the thawed portion weighed. Thaw drip was carried out in duplicate and calculated as the percent original weight lost by thawing. Mineral composition was determined in duplicate using the method of Teeny et al. (1984) by personnel from the UR Division in Seattle, Washington.
We used 7-point scales to evaluate color, flavor, chewiness, moistness, and desirability. The control sample was defined as having a score of 4 for color, chewiness, and moistness. A score greater than 4 indicated the experimental sample was rated as lighter in color, more tender, or more moist than the control. Scores lower than 4 indicated that the sample was darker, tougher, or drier than the control.
The control sample was defined as having a flavor score of 7, which was excellent fresh salmon flavor. A sample ballot is included as an appendix. Flavor ratings of 6 to 1 were defined as good but not intense fresh flavor, only mildly pleasant, trace of bad (rancid), mildly bad, strongly bad, and intensely bad, respectively. Thus, the flavor scale was partly hedonic rather than purely intensity in nature.
On the desirability scale, 7 was defined as like extremely; 6, like moderately; 5, like slightly; 4, neutral; 3, dislike slightly; 2, dislike moderately; 1, dislike extremely. A 5-point categorical scale was used to rate texture: 5, fibrous; 4, grainy or mealy; 3, flaky and firm; 2, soft; 1, mushy. Panelists were allowed to rate the texture and desirability of control samples along with the experimental samples.
No more than 4 samples were served at a seating of the panel so it was necessary to determine if grouping of samples affected the sensory scores. At 3 months of frozen storage, each block was tested twice. At the first seating of the panel, the judges were served fillet (0% mince) and 100% mince portions from once-frozen and reprocessed blocks. The next day, portions with 25% or 50% mince were served. These two seatings formed a grouping of the samples that emphasized the effect of reprocessing.
One week later, the panel was served portions from the once-frozen blocks containing 0 (fillet), 25, 50, or 100% mince one day and the analogous samples from the reprocessed blocks the next day. The second grouping therefore contrasted block form or the effect of mince.
An analysis of variance (ANOVA) treatment of the data indicated no significant interaction between the grouping effect and other experimental treatments. A lack of interaction meant that there was no significant effect with respect to which samples were served together. Furthermore, the mean panel scores were the same except for moistness scores. The scores for the 3-month samples were pooled. At most seatings, the panel was served portions with identical frozen storage histories but with different amounts of mince, thus contrasting block form.
After 1 month of frozen storage, malonaldehyde content ranged from 1.30 to 1.60 mg/100 g, indicating little rancidity (Table 3). After 3 months, the results were similar with the reprocessed blocks having lower values than the once-frozen blocks. This may have been caused by unintentional deep skinning of previously frozen fish. The tissue just under the skin of pink salmon has a high fat (lipid) content and should be more susceptible to oxidative rancidity. Leaving this fatty layer on the fillet was much easier with fish that had not been previously frozen. However, lipid content was not affected by reprocessing (data not shown). After 6 months, the malonaldehyde content of the samples was unchanged and after 12 months appeared to have decreased in some cases. These apparent decreases may be the result of variability within the blocks or the reaction of malonaldehyde with other components in the flesh. More importantly, although the chemical test for rancidity indicated little problem, these samples were ground before analysis so any rancidity that occurred on the surface of a block would have been diluted.
The Hunter a* value is a measure of how much red hue is in the color of a sample (the higher the a* value, the redder the sample). Addition of mince did not significantly affect the a* values at any given period of frozen storage (Tables 5 and 6). At 3 months frozen storage, reprocessing did not affect a* values, but it did at 6 months. The blocks from fish reprocessed at 6 months tended to have higher a* values (were more red) than once-frozen blocks or fish reprocessed at 3 months. At 12 months, reprocessing affected a* values but in a manner that varied to some degree with block form. Some of the differences between samples may have been caused by the natural variation in color of the samples.
The Hunter b* values were not affected by addition of mince at 1 month but tended to increase at 3, 6, and 12 months indicating a more yellow hue. This may reflect the greater susceptibility of minced flesh to changes during frozen storage. Reprocessing did not affect b* values at 3 or 12 months but, at 6 months frozen storage, the blocks made at 3 months tended to have lower b* values than the other blocks. As with the Hunter a* values, at least some of the differences between samples may be due to the inherent variability of the flesh of the fish.
Sensory scores (Table 7) at 1 month of frozen storage indicated that addition of mince did not significantly affect color, chewiness, or moistness scores but did affect flavor and desirability scores (Table 8). Addition of mince decreased the intensity of fresh salmon flavor but no off flavor was detected. Addition of mince significantly (Chi square G = 19.455, P = 0.004, df = 6) affected the distribution of texture scores causing an increase in grainy texture (Table 9). The samples with 50% or 100% mince had significantly lower desirability scores but all samples were of acceptable quality.
Color scores were not significantly affected by addition of mince but were affected by reprocessing as portions from once-frozen blocks fish tended to have lighter color than portions from reprocessed H&G fish (Tables 8 and 10). Flavor scores were not significantly affected by addition of mince but were affected by reprocessing as portions from once-frozen blocks tended to have more fresh salmon flavor.
Addition of mince did not significantly affect chewiness scores(Tables 8 and 10) but reprocessing did. Portions from reprocessed H&G fish were rated more chewy than portions from once-frozen blocks. The interaction for chewiness scores was significant because addition of mince made portions from once-frozen blocks chewier but did not affect portions made from reprocessed blocks since reprocessing itself had already toughened the flesh to some degree.
Moistness scores were significantly affected by addition of mince as increased mince was associated with increased dryness(Tables 8 and 10). The effect of reprocessing on moistness was not significant. Texture (Table 9) became more grainy with addition of mince (G = 59.381, P < 0.001, df = 9) but was not affected by reprocessing (G = 6.723, P = 0.081, df = 3). Desirability scores were lower for samples with added mince and those made from reprocessed H&G fish rather than unfrozen fish (Tables 8 and 10). However, all samples were of acceptable quality.
Color scores of portions from once-frozen blocks were usually higher indicating lighter color than portions from reprocessed blocks (Tables 11 and 12). Addition of mince did not affect color scores (Table 12). Flavor scores were generally lowered by addition of mince but were not affected by reprocessing(Tables 11 and 12). No sample had a mean flavor score lower than 4.2 which indicated little problem with rancidity.
Addition of mince increased chewiness of portions from the once-frozen blocks but not from reprocessed blocks (Tables 11 and Table 12). Reprocessing did not increase chewiness of portions except for blocks of fillets (F = 6.18, P= 0.006). Therefore, the overall effects of addition of mince and reprocessing were not significant, but the interaction between block form and reprocessing was significant (12).
Moistness scores were not affected either by addition of mince or reprocessing (Tables 11 and 12). Addition of mince increased graininess (G = 53.549, P < 0.001, df = 9) but reprocessing did not affect the distribution of texture scores (G = 6.474, P = 0.372, df = 6).
Mean desirability scores for the samples prepared from unfrozen fish decreased by more than 2 points with the addition of mince (Tables 11 and 12). For the samples prepared at 3 months and held an additional 3 months as blocks, a smaller decrease was seen. For the samples from blocks reprocessed at 6 months, the block with 25% mince scored higher than the other samples. Separate ANOVAs performed on data from each block form revealed that reprocessing affected desirability for the blocks with 0% mince (fillets) but not for the blocks with 25, 50, or 100% mince (Table 11). This difference explains the significant interaction between block form and reprocessing. Only two samples had mean scores less than 4.0, so the overall quality was still acceptable for all samples.
At 12 months of frozen storage, minor surface discoloration due to oxidation was seen in some of the blocks made at 0, 3, or 6 months. Usually, discoloration was seen in blocks from which samples had been taken earlier in the experiment and was probably the result of loss of packaging integrity. The samples were not trimmed to remove discolored material in order to realistically simulate possible commercial conditions where loss of packaging integrity could happen.
Mean color scores were lower (darker) for the 100% mince blocks than for the blocks containing fillets (Tables 13 and 14). The effect of mince on color scores was most obvious for the samples reprocessed at 6 or 12 months. Increased frozen storage as H&G fish prior to reprocessing into blocks also tended to decrease color scores, especially for the 100% mince blocks. Flavor scores at 12 months were affected by the amount of mince but not by storage time as H&G fish (Tables 13 and 14). The effect of mince on flavor was most evident for the blocks reprocessed at 6 and 12 months. In general, the flavor scores indicated little or no problem with oxidative rancidity.
Storage as H&G fish prior to being made into blocks did not affect chewiness scores but addition of mince made portions somewhat tougher, especially for once-frozen blocks (Tables 13 and 14). Although chewiness scores for reprocessed blocks did not appear to be as sensitive to addition of mince (Table 13), the interaction term was not significant (Table 14). Moistness scores followed the pattern of the chewiness scores with the fillet block from unfrozen fish rated as most moist and the minced samples being a little drier (Tables 13 and 14).
Chi-square analysis of texture scores (Table 9) at 12 months indicated that addition of mince continued to significantly affect results (G = 72.119, P < 0.001, df = 6) as blocks with 0% (fillets) or 25% mince generally had a flaky texture but some fibrousness was detected. Blocks with 50% or 100% mince usually had a grainy texture. Frozen storage as H&G fish prior to reprocessing affected the texture of blocks with 50% mince (G = 13.244, P = 0.039, df = 6) but not for the other blocks. For the blocks with 50% mince, the texture was usually rated grainy except for the block reprocessed at 6 months which was rated flaky. In general, the effect of reprocessing on texture scores was not significant (G = 3.972, P = 0.681, df = 6). Desirability scores were not affected by reprocessing but addition of mince decreased scores (Tables 13 and 14).
Flavor scores were affected by all experimental treatments, but no interaction term was significant(Tables 13 and 14). Addition of mince tended to lower flavor scores, especially at 6 and 12 months frozen storage. Once-frozen blocks tended to have higher flavor scores than reprocessed blocks. Flavor scores decreased with storage time for blocks made from either unfrozen or reprocessed fish, but only one sample had a mean value lower than 4.0 indicating that, in general, the samples were acceptable.
Chewiness scores (Table 15) were increased by addition of mince, reprocessing, and frozen storage (Table 16). The Repro x Form interaction term was significant because addition of mince resulted in tougher portions for once-frozen blocks but not for reprocessed blocks (Table 17).
Moistness scores (Table 15) were significantly affected by block form and storage time (Table 16). The portions were rated drier with increased amounts of mince and with increased storage time. The effect of reprocessing on moistness scores was not significant.
Addition of mince significantly affected texture scores (G = 130.65, P = 0.000, df = 9). The texture of blocks with 0% (fillets) or 25% mince was usually rated as flaky, while 50% and 100% mince blocks were rated as grainy (Table 18). Most judges indicated that the grainy texture was not unpleasant, just different from the fillet. The effect of reprocessing on texture scores was not significant (G = 7.532, P = 0.057, df = 3). Time of frozen storage did not affect texture scores for the blocks with 0, 50, or 100% mince but did for the 25% blocks (G = 15.516, P = 0.004, df = 4) where increased incidence of fibrous texture ratings occurred between 6 and 12 months. The reason for this difference is not apparent.
With the exception of the 100% mince block made from reprocessed H&G fish at 12 months, all samples had desirability scores of 3.5 or greater, indicating that their quality was still acceptable. Desirability scores (Table 15) indicated that portions from reprocessed blocks (H&G) were not as desirable as portions from once-frozen blocks (Table 16). Addition of more than 25% mince tended to lower desirability scores.
Determining a cause of the significant three-way interaction (Repro x Form x Form, Table 16) for desirability scores is best done by examining the Form x Stime interaction terms of samples stored as blocks or as H&G fish (Table 17). While the Form x Stime interaction was significant for reprocessed H&G samples, it was not significant for samples stored as blocks. This difference in behavior thus explains why the Repro x Form x Form interaction of Table 16 is significant. The significant interaction between storage time and block form for reprocessed blocks was probably caused by block form affecting desirability scores at 6 and 12 months but not at 3 months. Thus, the decrease in desirability due to addition of mince was smallest for material with frozen storage of 3 months or less or for material made from unfrozen fish. The decrease in desirability was greatest for material with total frozen storage time of 6 months or more or for material made from reprocessed H&G fish.
At 6 months of frozen storage, both addition of mince and reprocessing affected E values (Tables 19 and 20). As at 1 and 3 months, addition of mince tended to increase E values. The portions from blocks that had been made from H&G fish reprocessed after 3 months frozen storage tended to be tougher than the other comparative samples, a trend not clearly seen in the sensory data. The interaction between block form and reprocessing was significant because no change was seen with addition of mince in blocks made from unfrozen fish, whereas addition of mince tended to increase the E value for blocks made from reprocessed H&G fish (Table 19). No mechanical texture testing was performed on the 100% mince block made from unfrozen fish.
At 12 months of frozen storage, addition of mince increased E values in most but not all cases (Tables 19 and 20). No reason was evident to explain why the 50% mince blocks had a higher E value than the 100% mince blocks for samples reprocessed at 3 and 12 months of frozen storage. Storing the fish in H&G form prior to being made into blocks may have decreased E values in the case of the blocks with no mince, but the opposite trend was seen for blocks with mince. Evidently, minimizing the time minced salmon spends in frozen storage reduces toughening.
To determine the effect of frozen storage on E values, separate ANOVAs were performed on data from blocks made from unfrozen fish (H&G time = 0) and blocks examined immediately after being made from previously frozen H&G fish (H&G time = 3, 6, or 12). For samples prepared from unfrozen fish, the E values increased during frozen storage especially between 6 and 12 months and for samples containing mince (Tables 21 and 22).. For samples made from reprocessed fish, the effect of storage time was significant for the 0% (fillets), 25%, and 50% blocks but not for samples from blocks containing 100% mince (Table 21).
The E values were significantly (P < 0.001) correlated with mean chewiness and desirability scores (r = -0.569 and -0.621, respectively). The negative coefficients meant that samples with high E values tended to be chewier and less desirable.
Although the quality of blocks made from unfrozen fish was generally better than blocks made from reprocessed H&G fish, portions made from reprocessed fish were acceptable. According to the results of sensory analysis, frozen storage produced a slight but significant darker color for blocks made from reprocessed H&G fish. Decreases in flavor scores during frozen storage were minor and attributed more to the loss of fresh flavor than the occurrence of rancidity. Toughness as measured by chewiness scores increased during frozen storage for blocks made from unfrozen fish but not for blocks made from reprocessed H&G fish. Thaw drip may be a problem for reprocessed material as frozen storage time approaches 12 months. Results of mechanical textural analysis agreed with the trends observed by sensory analysis.
Bourne, M. C. 1982.
Food texture and viscosity. Academic Press, New York, 325 p.
Horowitz, W. (editor). 1980.
Official methods of analysis, 13th edition, p. 215. Association of Official Analytical Chemists, Washington, DC.
Lemon, D. W. 1975.
An improved TBA test for rancidity. Canadian Fisheries and Marine Service New Series Circular, No.51, 4 p.
Nie, N. H. 1975.
SPSS, Statistical Package for the Social Sciences. McGraw-Hill, New York, 988 p.
Pacific Associates. 1994.
Recovering world leadership in salmon. Prepared for the Alaska Department of Commerce and Economic Development, P. O. Box 110800, Juneau, AK 99811-0800, 70 p.
Reppond, K. D., J. Collins, and D. Markey. 1985.
Walleye pollock (Theragra chalcogramma): Changes in quality when held in ice, slush-ice, refrigerated seawater and CO2 modified seawater then stored as blocks of fillets at -18°C. J. Food Sci. 50:985-989 and 996.
Sokal, R. F., and F. J. Rohlf. 1969.
Biometry. Freeman, San Francisco, 776 p.
Teeny, F. M., E. J. Gauglitz, Jr., A. S. Hall, and C. R. Houle. 1984.
Mineral composition of the edible muscle tissue of seven species of fish from the Northeast Pacific. J. Agric. Food Chem. 32:852-855.
Table of Contents