SHELF-LIFE OF FRESH FILLED PASTA. HAZARD ANALYSIS AND CRITICAL CONTROL POINTS OF THE MANUFACTURING PROCESS AND HOUSEHOLD PRACTICES

Hazard Analysis and Critical Control Point (HACCP) approach was used to assess the safety of a fresh filling pasta (ricotta-filled ravioli) manufacturing plant in La Plata region, Argentina. Household practices like cooking and holding meals before serving were also evaluated. Samples of ricotta, main raw material of the filling, showed colony counts of Enterobacteriaceae total microorganisms, molds and yeasts of (5 × 10³-1 × 10⁵), (1x10⁵−1x10⁸) and 1 × 10⁵ CFU/g, respectively. E. coli was also detected in one out of five samples, suggesting a hazardous condition of this raw material. Salmonella spp. was not isolated from any of the dough samples tested. Ricotta filling showed high colony counts of mesophilic microorganisms (6 × 10⁶ CFU/g), Enterobacteriaceae (1 × 10⁵ - 1 × 10⁶ CFU/g), while E. Coli was detected in 20% of the samples. Counts of B. cereus, S. aureus were less than 1 × 10² CFU/g in the analyzed materials. In the finished product (ricotta-filled ravioli), the colony counts of mesophilic microorganisms and Enterobacteriaceae were 3 × 10⁸ and 8 × 10⁵ CFU/g, respectively. Critical Control Points found were cooking and holding time before serving. The implementation of suggested corrections allowed the microbial quality of the final product (ricotta-filled ravioli) to be improved considerably. The growth of total microbial counts during refrigerated storage (0, 4, 8 and 10C) were measured and the shelf-life of ricotta and ricotta-filled ravioli was calculated.     

EXAMINATION OF THE MICROBIAL STATUS OF SELECTED INDIGENOUS FERMENTED FOODS IN NIGERIA

Four Nigerian traditionally fermented foods (wara, nono, ogi and kununzaki) were evaluated for the presence of some microorganisms of public health concern. Among the dairy foods , Staphylococcus aureus and Klebsiella sp. were isolated from wara while Escherichia coli, Salmonella sp. and Klebsiella sp. were isolated from nono. The cereal-based fermented foods (ogi and kunu-zaki) contained Bacillus subtilis, E. coli, S. aureus, Klebsiella sp. and Enterococcus faecalis. The mesophilic aerobic counts were: 5 × 10⁵ for wara; nono, 1.53 × 10⁷; ogi, 3.6 × 10⁶ and kunu-zaki, 2.6 × 10⁶ cfu/mL. The enterobacteriaceae counts on nono, wara, ogi and kunu-zaki were 1.79 × 10⁷, 4.5 × 10⁵, 4.0 × 10⁵ and 1.2 × 10⁶ cfu/mL, respectively. No Vibrio count (detection limit: <10 cfu/mL) was recorded in all the food samples considered. The yeast and mold counts ranged from 1.0 × 10⁵– 3.31 × 10⁷ among the food products. The antimicrobial susceptibility patterns of the organisms isolated from dairy products (nono and wara) revealed that they were resistant to ampicillin (100%) and sensitive to gentamicin (100%) and nalidixic acid (100%). Most isolates from cereal based products (ogi and kunu-zaki) were 100% resistant to penicillin, ampicillin and chloramphenicol. This work highlights the need to maintain hygienic standards in the preparation of our locally fermented cereal and dairy foods.     

HYGIENIZATION OF INDIAN CHICKEN MEAT BY IONIZING RADIATION

Both fresh and frozen chicken meat were evaluated for microbiological status by screening for total bacterial counts and for the presence of pathogens like Enterobacteria , Bacillus cereus, coagulase positive Staphylococci and Salmonella spp. Most of the samples exhibited heavy bacterial contamination (1.2 × 10⁵ - 2.6 × 10⁶/g), mainly with Staphylococcus spp. (1.5 × 10⁴ - 2.8 × 10⁵/g). All the chicken samples also showed the presence of Salmonellae (3 × 10¹ - 2.1 × 10²/g). Among the different serotypes observed in chickens . S. typhimurium was common in fresh as well as frozen chicken. Radicidation at 2 kGy at cryogenic conditions (−40°C) was efficient in eliminating the natural pathogenic contamination of the poultry . Salmonella spp. viz. S. seftenberg and S. typhimurium differed in radiation sensitivity, the D₁₀ values in phosphate buffer (pH 7.2) being 0.25 kGy and 0.12 kGy, respectively. Chicken homogenate (10%) offered approximately 2-fold protection to these cells. Chicken samples artificially inoculated with a heavy inoculum (10⁸ cells/g) of these 2 serotypes required higher gamma radiation doses of 4–5 kGy. The findings suggested that a dose of 2 kGy is adequate for normally contaminated chicken samples, but for the heavily contaminated chicken a dose of 4–5 kGy, depending upon the predominating Salmonella serotype present, is required .     

Rapid dye reduction tests for the determination of microbiological quality of meat

Rapid dye reduction tests have been developed to determine the quality of meat. Three chemical indicators, resazurin and two tetrazolium compounds, were used to correlate the microbial numbers and reduction times in meat samples. Twenty-five surface samples from sheep carcasses were subjected to each reduction test. Total viable counts given were obtained at 37°C. Resazurin reduction time was 90–120 min when the bacterial counts ranged from 1.5×10⁶ to 7.7×10⁶/cm². Samples showing bacterial counts between 1.5×10⁶ and 6.0×10⁶/cm² reduced tetrazolium (NBT) in 360–390 min whereas samples containing bacterial counts of 2.1×10⁶/cm² took 420–450 min to reduce iodophenyl nitrophenyl tetrazolium (INT) dye. Regression equations relating the number of organisms per cm² and reduction time were applied to predict the microbiological quality of meat samples from reduction time data. Among the three dyes, resazurin gave the lowest reduction time.     

EVALUATION OF BACTERIOLOGICAL QUALITY IN SELECTED COMMERCIAL INFANT FORMULAS AVAILABLE IN POLAND AND EGYPT

One hundred samples of commercial infant formula bought in shops in the Poznan region (Poland) and Cairo region (Egypt) were investigated. Samples were analyzed for aerobic plate counts (APC), total Bacillus cereus (TBC), and incidences of Bacillus spp. and coliforms. The mean APC and TBC did not show any important variation with country, being practically the same in products bought in Poland and Egypt. All commercial infant formula analyzed immediately after opening were of satisfactory bacteriological quality, exhibiting APC lower than 10⁴ CFU g⁻¹ (mean 4.9 × 10²) and TBC lower than 10³ CFU g⁻¹ (mean 1.1 × 10²). However, 60% of the examined fruit juice and ready-to-feed infant formula presented TBC above the recommendation safety limit after storage at 22C for 72 h and at 35C for 48 h. In most cases the mean log APC and TBC were highest (P > 0.05) for fruit juice and ready-to-feed infant formula stored at elevated temperature (35 ± 1C).     

THE EFFECT OF AEROBIC MESOPHILIC MICROFLORAL LEVELS ON THE ISOLATION OF INOCULATED LISTERIA MONOCYTOGENES STRAIN LM82 FROM SELECTED FOODS

The effect of aerobic mesophilic microfloral concentration on the isolation of Listeria monocytogenes LM82 was studied in 31 (18 cheeses and 7 noncheese) retail foods having standard plate counts of 10¹ to 10⁸ colony forming units (CFU)/g. Foods were spiked with L. monocytogenes and enriched at 30°C for 24 h in a selective enrichment broth used in a U.S. Food and Drug Administration method. Inoculum levels for isolation on modified McBride agar ranged from 0.1 to > 5 × 10³ with a geometric mean value of 5 inoculated CFU/g or 1.4 CFU/g. Pure Enterococcus (Streptococcus) faecalis ( 0 to 6 × 10⁶ inoculated CFU/mL ) in the absence of food matrix had no effect on the enrichment of L. monocytogenes. Ease of isolation of LM82 was independent of the food microflora concentration both generally and in the specific food type of 9 Brie cheeses. Competition, when it occurs, therefore, may be due to specific bacterial competitors rather than bacterial numbers .     

QUANTITATIVE STUDIES OF PROTEOLYTIC AND LIPOLYTIC BACTERIA IN FROZEN MINCED BEEF AND HAMBURGER

Sixty frozen samples (30 hamburger patties and 30 minced meat) purchased from different retail markets in Ismailia, Egypt were examined for incidence of proteolytic and lipolytic bacteria. Mean values of proteolytic psychrophiles in the examined samples of hamburger and minced meat were 2×10⁵ and 6×10⁴, respectively. Lipolytic psychrophiles in hamburger were 8×10⁵ and 3×10⁵ in minced meat. Proteolytic mesophiles in hamburger and minced meat were 6×10⁵ and 5×10⁴, respectively. Mean values of lipolytic mesophiles were 5×10⁴ for hamburger and 5×10³ for minced meat. Proteolytic thermophiles in hamburger and minced meat were 3×10³ and 10³, respectively. Both proteolytic and lipolytic activities were exhibited by various bacteria that were identified.     

REMOVAL OF SALMONELLA TYPHIMURIUM ATTACHED TO CHICKEN SKIN BY RINSING WITH TRISODIUM PHOSPHATE SOLUTION: SCANNING ELECTRON MICROSCOPIC EXAMINATION

The effect of trisodium phosphate (TSP) on Salmonella typhimurium attached to chicken skin was investigated by using scanning electron microscopy (SEM). Chicken drumsticks were inoculated with Salmonella typhimurium (2 × 10⁸ CFU/mL) for 30 min. Both inoculated and non-inoculated drumsticks were rinsed with 10% TSP solution at 10 or 50C for 15 s, and skin pieces were cut and fixed for SEM examination. For inoculated skins, a significant difference was noticed between TSP-rinsed and control skins (water-rinsed) at both temperatures. While control skins were covered with salmonellae (4 × 10⁵∼ 1 × 10⁶ CFU/cm²) and miscellaneous debris, TSP-rinsed skins, either at 10 or 50C, showed clean skin surfaces (<8×10³ CFU/cm²). For non-inoculated skins, it was difficult to see the difference in the number of attached bacteria due to their low numbers, however, water-rinsed skins still showed the debris on the surface. Above observations suggest that one of the major mechanisms of TSP on salmonellae reduction is detachment of contaminants from the skin surface.     

INHIBITION OF STAPHYLOCOCCUS AUREUS AND BACILLUS CEREUS ON A VEGETARIAN FOOD TREATED WITH NISIN COMBINED WITH EITHER POTASSIUM SORBATE OR SODIUM BENZOATE

Various amounts of nisin (0, 10³ and 5 × 10³ IU/g) in combination with either potassium sorbate (0, 2, and 3%) or sodium benzoate (0, 0.06 and 0.12%) were tested for effectiveness in inhibiting growth of Staphylococcus aureus C10 and Bacillus cereus B7 inoculated on a vegetarian food. The strains used were isolated from vegetarian foods obtained commercially in Taiwan, and the test food, spice and dried bean curd, was selected for the study based on ability to support the growth of these organisms. After treatment with a preservative combination, the surfaces of sterilized food samples were inoculated, samples were stored in vacuum or nonvacuum packages at either 4C or 30C, and at appropriate times, tested for microbial growth. Growth of both isolates was unaffected by vacuum-packaging treatment; however, a bacteriostatic effect was found at 4C. Data indicated that during the 14-day storage at 4C, vacuum-packaged samples treated with 5 × 10³ IU/g nisin and 0.12% sodium benzoate significantly (p < 0.05) decreased the counts of S. aureus C10 and B. cereus B7 by 2.61 and 3.02 log₁₀ CFU/g, respectively. In the vacuum-packaged samples treated with 5 × 10³ IU/g nisin and 3% potassium sorbate, counts for C10 and B7 were decreased by 2.35 and 2.64 log₁₀ CFU/g, respectively. Thus, the combined treatment extended the shelf-life of the vegetarian food .     

MICROBIOLOGICAL QUALITY OF TRADITIONAL MARKET CURED FISH IN TANZANIA

The microbiological quality of six varieties of retail market traditionally cured fish in Morogoro, Tanzania was investigated over a five-month period. The fish were contaminated with bacteria and molds at levels of: total aerobes, 10⁶ - 1.7 × 10⁷ c.f.u/g; faecal coliforms, 1.1 × 10¹ - 2.5 × 10³ MPN/g; faecal streptococci, 1.4 × 10¹ - 1.3 × 10³ MPN/g; Staphylococcus aureus, 1.3 × 10³ - 8.6 × 10³ c.f.u/g; Aspergillus flavus group, 2.1 × 10¹ - 2 × 10² c.f.u/g of fish. Of faecal coliform, 45% of the isolates were Escherichia coli. Twenty-five percent of the S. aureus isolates were coagulase positive. Sixteen percent of A. flavus isolates were aflatoxigenic. Aflatoxin contamination ranged from O to 18.5 μg/kg of fish. Insect infestation by Dermestes spp. and mites was observed. The results of this preliminary study emphasize the importance of proper processing and handling offish in the tropics in order to safeguard public health .     

The relationship of bacterial numbers and types to diamine concentration in fresh and aerobically stored beef, pork and lamb

Bacterial numbers and putrescine and cadaverine concentrations were measured in intact beef, pork and lamb and minced beef at retail and during aerobic chill storage at 5°C. Putrescine concentrations increased consistently with 'total' aerobic viable count (TAVC) but cadaverine concentrations increased only when high numbers of presumptive Enterobacteriaceae were present. Significant changes in diamine concentration did not occur until the TAVC exceeded 4.2 × 10⁷/cm² or g when the meat was clearly spoiled. Changes prior to the onset of spoilage were not sufficient for their use as a predictive indicator of the acceptability of the meat.     

Decontamination Efficacy of Combined Chlorine Dioxide with Ultrasonication on Apples and Lettuce

ABSTRACT:  The bacterial reduction of Salmonella and Escherichia coli O157:H7-inoculated apples and lettuce by ClO₂ at 0, 5, 10, 20, and 40 ppm with and without 170-kHz ultrasonic treatments for 3, 6, and 10 min, respectively, have been studied. The treatments of ClO₂ at 20 and 40 ppm for 3, 6, and 10 min or at 5 and 10 ppm for 6 and 10 min with 170-kHz ultrasonication caused 3.115 to 4.253 log reductions in Salmonella and 2.235 to 3.865 log reduction in E. coli O157:H7 on inoculated apples. Using combined ClO₂ and ultrasonication to treat 4.48 × 10⁴ CFU/g Salmonella and 1.07 × 10⁵ CFU/g E. coli O157:H7-inoculated lettuce, the bacterial reductions were 2.257 to 2.972 and 1.357 to 2.264 log, respectively. The residual ClO₂ decreased with increasing treatment times, over 80% of ClO₂ was detected after the 3-min treatment, and more than 70% remained after the 10-min treatment time. No bacteria were recovered from the posttreatment solutions of ClO₂ or ClO₂ combined with ultrasonication. The temperature of the ClO₂ treatment was 20.1 °C, and it increased to 40.1, 44.9, and 50.3 °C, with 170-kHz ultrasonic treatments for 3, 6, and 10 min, respectively, on apples.     

EFFECT OF HIGH HYDROSTATIC PRESSURE, GAMMA-IRRADIATION AND COMBINATION TREATMENTS ON THE MICROBIOLOGICAL QUALITY OF LAMB MEAT DURING CHILLED STORAGE

The effect of gamma irradiation (1.0 kGy) and high hydrostatic pressure (200 MPa for 30 min), either alone or in combination on the shelf-life of lamb mince meat at 0–3C was studied. Untreated control samples initially had total microbial counts of 10⁵ CFU/g, 10² CFU/g of coliforms and 10⁴ CFU/g of Staphylococcus spp. Coliforms were eliminated by all the treatments . Staphylococcus spp. however, were reduced only by 1 log cycle when treated with irradiation alone and high pressure alone. These species were a mixture of mannitol-fermenting and mannitol-nonfermenting strains. In samples subjected to the combination treatment , Staphylococcus spp. appeared only after 3 weeks of storage and all were mannitol-nonfermenting. On the basis of microbiological and sensory quality, the shelf-life of the control sample was less than 1 week. All treated meat samples had a shelf-life of 3 weeks, but only combination treated samples were free from potentially pathogenic Staphylococcus spp .     

Spectroscopic Differentiation and Quantification of Microorganisms in Apple Juice

ABSTRACT: A fast and easy-to-operate Fourier Transform Infrared (FTIR) spectrometry-based approach was developed for microbial differentiation and quantification in apple juice. Eight different microorganisms were evaluated: Enterobacter cloacae, Salmonella typhimurium, Enterobacteraerogenes, Salmonella choleraesuis, Serratia marcescens, Pseudomonas vulgaris, Vibrio cholerae , and Hafnia alvei . FTIR spectroscopy combined with chemometrics could differentiate the microorganisms studied at low concentration level of 10³ colony-forming units (CFU) /mL in apple juice. The chemometric models developed to count microorganisms in apple juice were validated by an independent test set consisting of 18 samples and correlated against plate counts satisfactorily up to a detection limit of 10³ CFU/mL.     

MICROBIOLOGICAL QUALITY OF THE DAIRY PRODUCT PEDHA AND ITS IMPROVEMENT USING GAMMA RADIATION

Eighteen pedha samples procured from A and B grade retail shops were examined for their overall microbiological quality and for the presence of foodborne pathogens viz. Staphylococcus aureus, Salmonella sp., Coliforms , Listeria monocytogenes, Yersinia enterocolitica and Bacillus cereus. The microbiological quality of pedha samples from B grade shops was very poor as compared to pedha from A grade shop as evidenced by the very high total bacterial counts (6 × 10⁷ cfu/g), high counts of S. aureus (as high as 7 × 10⁶ cfu/g) and presence of coliforms and Listeria and Yersinia sp. in 33% of the samples. All the samples from A grade shops were also positive for S. aureus though negative for coliforms , Yersinia, Salmonella, Listeria and B. cereus. Gamma irradiation of pedha at a dose of 3kGy at 0C reduced overall bacterial load by five log cycles and S. aureus and coliforms could be totally eliminated. However, 5 kGy dose was necessary to eliminate S. aureus if the initial number exceed 1 × 10⁵ cfu/g. Inoculated pack studies confirmed that 3 kGy dose was sufficient for the complete elimination of up to 1 × 10⁵ cfu/g of S. aureus. A dose of 3 kGy had minimal effect on the sensory quality of pedha and even pedha samples irradiated at 5 kGy dose were acceptable. Treatment with 3 kGy dose of gamma radiation totally eliminated S. aureus and coliforms in pedha, thereby ensuring their microbial safety.     

RADIATION STERILIZATION OF SPICES FOR HOSPITAL FOOD SERVICES AND PATIENT CARE

A survey of the commercial spices used by food services in a typical hospital environment revealed high contamination with microorganisms, i.e., 10⁴ to 10⁷ counts per gram. The predominant microorganisms were as followed (in colony counts/gram): (1) heat-resistant bacterial spores in black pepper, 1 × 10⁷; thyme, 2 × 10⁶; anise, 7 × 10⁴; curry powder, 4 × 10⁵; poultry seasoning, 8 × 10⁴; pickling spice, cardamom, and cumin, 1.5–3 × 10⁴; (2) mixed populations of vegetative cells and bacterial spores in cumin, 1 × 10⁶; (3) molds in cream of tartar, 2 × 10⁴. Sterility of food may be important in a hospital setting, especially in the care of immunocompromised patients. To eliminate the organisms, we recommend radiation treatment, accompanied by appropriate microbiological quality control. On the basis of radiation survival data, the composite natural flora would be reduced to the level of "commercial sterility" (defined as less than 10 organisms per gram((Kiss 1982) by the following minimum radiation doses (in kGy): black pepper, 13; thyme, 13; cumin, 12; anise, 10; curry, 7.3; pickling spice, 7; poultry seasoning, 6; cardamom, 9.4; cream of tartar, 4. For practical purposes, two dose levels can be recommended for treatment of spices in the hospital environment, low = 6–10 kGy and high = 10–15 kGy.     

Enhancement of Fermentation Process in Pu-Erh Tea by Tea-Leaf Extract

ABSTRACT:  Pu-erh tea is known as a fermented tea and longer storage enhances its flavor and taste. Recently,  Aspergillu s,  Blastobotrys , and  Streptomyces  are found to play important roles in nutritional enhancement of Pu-erh tea by fermentation. Since water and temperature affect the microbial growth, we therefore explored the factors that might enhance the Pu-erh tea fermentation. The results showed that the addition of fresh tea-leaf extract (TLE) enhanced the withered tea fermentation (at 37 °C, 80 to 85% RH) as compared with the water only. Contents of statin, GABA, gallic acid, DPPH scavenging and polyphenol oxidase (PPO) activities were increased, whereas polyphenols and caffeine were decreased over 6 mo. TLE dose-dependently enhanced some of the qualities (that is, statin, PPO) of Pu-erh tea significantly as compared with the water only. The effect was related to the increase population of  A. niger  and  A. carbonarius  at 6 mo (from 7.6 ± 1.2 × 10¹ and 3.2 ± 1.3 × 10¹ to 3.1 ± 1.2 × 10⁶ and 2.4 ± 1.1 × 10⁵ colony forming units [CFU]/g, respectively). After drying process (90 °C, 30 min), the total microbial count from these samples returned to background level (3 ± 0.5 × 10² CFU/g). None of ochratoxin and fumonisin, toxins from  Aspergillus , was detected in the final products. The flavor and taste were also enhanced by treatment with TLE. The inoculation with  S. cinereus  Y11 with 2% TLE further enhanced these functional contents (about 2-fold increase of statin level) in the experimental Pu-erh tea. Therefore, this result may add a new process for Pu-erh tea manufacture.     

LINKING MICROBIOLOGICAL CRITERIA FOR FOODS WITH QUANTITATIVE RISK ASSESSMENT

The current annual risk of acquiring a foodborne disease in the United States is estimated at 2.7 × 10^-2. The risk of associated death is estimated at 3.7 × 10^-5. These represent a health care burden >$3 billion. Using a risk assessment model one can identify levels of microbial contamination which may be unacceptable in foods and appropriate controls needed to reduce these levels. Salmonella bacteria continue to represent a large percentage of the identifiable infections. A model developed from human dose-response studies predicts the probability of infection for Salmonella at 7.5 × 10^-3 with exposure to a single CFU of the organism. Risks of severity (hospitalization), mortality, reactive arthritides, and mortality in the elderly are estimated at 3.1 × 10^-6, 7.5 × 10^-6, 1.7 × 10^-5, and 2.8 × 10^-4. Exposure to microbial contaminants needs to be evaluated on a single meal basis. For chicken, exposure may range from a single drum stick (38g) to a half broiler (176g) but averages around 80g. For beef between 51 and 85g may be consumed during a single meal. Therefore, methods for monitoring must be able to detect at least 1 CFU/80g. Risks for some pathogenic E. coli are estimated at 1,000 to 10,000 less than Salmonella. Therefore, use of coliforms as indicators needs to be assessed and related to occurrence and survival and regrowth potential of the enteric bacteria of greater public health concern. Because, 20% of the U.S. population may be considered to be in a special population category and at an increased risk of severe outcomes, no more than 20% failure of a standard should be acceptable .     

LACTIC ACID INACTIVATION OF SALMONELLA TYPHIMURIUM ATTACHED TO CATFISH SKIN

This study was conducted to test sanitizer efficacy of lactic acid on Salmonella typhimurium populations that were firmly attached or loosely attached on catfish skin with or without mucus. For both cell types, counts of S. typhimurium were reduced from 5.9 log¹⁰ CFU/skin to below the limit of detection of 2.0 log¹⁰ CFU/skin by a 5 min exposure to 2% lactic acid. Exposure for 1 min reduced a 10⁴ CFU/skin inoculum to below the limit of detection. Catfish skin mucus slightly decreased the antimicrobial effect of lactic acid against firmly attached S. typhimurium.     

SPOILAGE AND HISTAMINE IN WHOLE PACIFIC HERRING (CLUPEA HARENGUS PALLASI) AND PINK SALMON (ONCORHYNCHUS GORBUSCHA) FILLETS

Fresh pink salmon (Oncorhynchus gorbuscha) fillets and whole Pacific herring (Clupea harengus pallasi) were stored for two weeks at 10C to determine if significant amounts of histamine were produced before the fish spoiled. Spoilage odors in salmon were moderate by day 4 and intense by day 7, while herring had detectable spoilage by day 4 and became potent by day 6. Aerobic bacterial counts increased from 4.0 × 10²/g initially to 3.6 × 10⁸/g in salmon fillets by day 14 and from 2.3 × 10³/g initially to 2.7 × 10⁷/g in whole herring by day 14. Total volatile nitrogen increased from 1.8 to 78.5 mg N/100 g in salmon and 2.2 to 23.6 mg N/100 g in herring. Histamine was not detected in salmon, while concentrations reached 54.9 ppm in herring at day 14. However, herring were considered spoiled by day 6.