Examination of Market Foods for

APPuED MICROmoLOGY, Sept. 1967, p. 1062-1069 Copyright 1967 American Society for Microbiology Vol. 15, No. 5 Printed in U.S.A. Examination of Market Foods for Coliform Organisms HERBERT E. HALL, DAVID F. BROWN, AND KEITH H. LEWIS Milk and Food Research Section, National Center for Urban and Industrial Health, Public Health Service, Cincinnati, Ohio 45226 Received for publication 16 March 1967 Food specimens (490) in nine categories were examined for total aerobic plate count and numbers and types of coliform organisms, including the enteropathogenic Escherichia coli (EEC). The total counts were compared with various suggested standards, and a limit of 100,000/g appeared to be a realistic goal, except for certain food types with a high level of natural flora. Plate counts in VRB were compared to counts obtained by isolation by enrichment in LST Broth, and the latter method produced a higher percentage of isolations. The presence of E. coli was determined by use of EC Medium incubated at 44.5 + 0.1 C. Only 40.4% of the positive EC tubes, however, contained E. coli. It appeared that a limit of 10 coliform organisms per g as a suggested standard could be met with several types of foods. Isolation of EEC was obtained only three times, or in 0.6% of the specimens. Most workers who examine food of various types (raw, frozen, or prepared for purposes of survey, product control, or the determination of the etiology of foodborne disease) accept the fact that some level of coliform contamination may be present (12). The exact significance of the presence of these organisms, however, is not clearly defined. In recommended standards for various foods, the allowable numbers range from none in infant food formulas (2) to 90/g in grade A dry milk (supplement 1 to Milk Ordinance and Code of 1953, Public Health Service, 1959) with, most of the recommendations centering around levels of 0/0.1 g (10) to 20/g (17; Grade A Pasteurized Milk Ordinance, Public Health Service, 1965). Other suggested standards require the determination of numbers of fecal coliform organisms, as compared to the number of bacteria capable of giving a positive reaction when grown in most probable number (MPN) tubes containing lactose, Lauryl Sulfate Tryptose (LST) Broth (BBL or Difco), or Brilliant Green Bile (BGLB) Broth, 2% (BBL or Difco) (National Shellfish Sanitation Workshop, 1964). The importance of the presence of enteropathogenic Escherichia coli (EEC) in foods has rarely been considered in this country, except when all environmental factors are being studied in outbreaks of infantile diarrhea. In the literature from Europe and Japan, however, incidents of foodborne disease caused by EEC are frequently reported (3, 18). The study reported in this paper was designed to determine the frequency of occurrence of EEC in market foods and to evaluate methods of detection and quantification of total coliform organisms and of E. coli. MATERIALS AND METHODS Food specimens. All foods used in this study were selected from the shelves or self-service cold boxes of various chain stores in the Cincinnati, Ohio, area or were purchased over-the-counter or from vending machines. No attempts at selection were made other than to determine that the wrapping materials on prewrapped goods were intact. The types of foods and the number of specimens of each type examined were: frozen ready-to-eat desserts, 23; cheese and cheese products, 26; baby foods in jars, 33; frozen fish and seafoods, 36; dry cereals and mixes, 46; raw and frozen vegetables, 49; prepared and convenience foods, 81; raw meats, 88; and prewrapped sandwiches, 108. Media. Commercially available dehydrated media were used throughout the study and prepared according to the manufacturers directions with the exception of the LST broth, which was used in 30-ml volumes at 1.33 strength to yield single-strength composition when a 10-ml volume of inoculum was added. Procedure. Of each food specimen, 50 g was removed aseptically from the original package, placed in a Waring Blendor jar (either glass or metal, 1-liter capacity) with 450 ml of buffered dilution water (2), and blended, at the speed of the "low" switch position (approximately 8,000 rev/min) for 2 min. Further dilutions of the food (10" through 10-7) were prepared in buffered dilution water (2). Total aerobic plate counts were obtained by plating 1-ml volumes of dilutions of 10-1 through 107 in duplicate in Plate Count Agar (Difco or BBL), incubating for 48 hr at 1062

VOL. 15, 1967 EXAMINATION OF MARKET FOODS FOR COLIFORMS 1063 35 C, and counting the colonies in duplicate plates containing between 30 and 300 colonies. Coliform counts were obtained by plating 10 ml of the 1:10 blend in three plates and 1-ml volumes of dilutions of 10-' through 10-' in duplicate in Violet Red Bile (VRB) Agar (BBL or Difco). Plates were poured with approximately 15 ml of VRB agar and overlaid with 4 to 5 ml of the same agar. The plates were incubated for 24 hr at 35 C, and all of the red colonies in the 10-ml plates or dilution plates containing 150 or fewer colonies were counted; 10 colonies were picked from the plates counted, or, when less than 10 colonies were present in the 10-ml plates, all colonies were picked. When both typical (colonies >0.5 mm in diameter, with a zone of precipitated bile) and atypical colonies were present, some of each were picked. The colonies were inoculated into tubes of BGLB Broth, incubated for 48 hr at 35 C, and observed for gas formation at 24 and 48 hr. All tubes showing gas formation were subcultured by streaking onto Eosin Methylene Blue (EMB) Agar (BBL or Difco). Tubes of EC Medium (BBL or Difco) were also inoculated from the positive BGLB tubes, and incubated for 48 hr at 44.5 (40.1) C in a thermostatically controlled water bath; those showing gas formation at 24 or 48 hr were subcultured to EMB. A 10-ml volume of the 1:10 dilution was added to a tube containing 30 ml of LST Broth at 1.33 strength. The tube was incubated for 48 hr at 35 C and, when gas was produced in 24 or 48 hr, it was subcultured to BGLB and EC broth. All positive BGLB and EC tubes were streaked on EMB Agar. All EMB plates were incubated for 24 hr at 35 C and examined for types of colonies present. Examples of each type on plates inoculated from both BGLB and EC were transferred to Veal Infusion Agar slants (Difco). The growth on these slants was used to determine the types of coliforms present by inoculating MR-VP medium broth (BBL), Simmons Citrate Agar (BBL), and Tryptose Broth (Difco) to obtain IMViC patterns. All strains giving patterns indicative of the E. coli types (++-- or -+--) were tested with antisera for EEC by the methods recommended by Edwards and Ewing (4). TABLE 1. RESULTS The total aerobic counts obtained with the various types of food ranged from no organisms detected in 1 ml of 10-' dilution to over 100 million organisms per g. The ranges, mean, and median counts are shown in Table 1. Consistently low counts (median <100 per g) were obtained in baby foods in jars; of the 33 specimens, 18 failed to show any growth at the 10-1 dilution. The highest count in baby foods (110/g) was obtained from a specimen of creamed corn; the next highest (75/g) was in vegetables with bacon. Most of the growth obtained from these specimens was of aerobic spore-forming bacilli; no attempts were made to determine the species. The highest mean and median counts were obtained with cheese and cheese products ( >10 million per g); the highest count obtained in the study, however, was a specimen of giblet gravy in the prepared and convenience food category. Because mean values tend to be overinfluenced by the occurrence of one or two very high values, the median counts give a more realistic picture of the microbial levels in these foods. The median values indicated that most of the foods examined had relatively low counts (Table 1). Cheese and cheese products, as well as sandwiches made with cheese, had high levels of residual organisms related to cheese production; raw meats and raw and frozen vegetables also had relatively high residual or natural levels of bacterial contamination. Frozen desserts, dry cereals and mixes, frozen fish and seafoods, and most prepared and convenience foods had relatively low levels (a few thousand per g) due to effects of processing or terminal heat treatment. Recommended standards of 50,000 to 100,000 organisms per g have been suggested for frozen precooked foods, shellfish and some milk products (11, 13; Quartermaster Food and Container Institute of the Armed Forces, MIL-M-13966, 1955). Of the Aerobic plate counts Class of food sneomeof Range Mean Median Frozen desserts... 23 0a_4.5 X 104 3.5 X 103 5.9 X 102 Cheese and cheese products... 26 0-5 X 107 2.1 X 107 1.7 X 107 Baby foods (jars)... 33 0-1.1 X 102 < 100/g <100/g Fish and seafoods.36 4.0 X 10O-4.0 X 105 7.6 X 104 7.0 X 103 Dry cereals and mixes... 46 1.2 X 106 5.6 X 103 1.7 X 102 Raw and frozen vegetables... 49 1.2 X 102-4.2 X 107 3.3 X 106 4.0 X 104 Prepared and convenience foods... 81 0-1.2 X 108 2.7 X 106 3.3 X 103 Raw meats... 88 1.5 X 10L3.2 X 107 2.5 X 106 2.8 X 105 Sandwiches... 108 0-3.0 X 107 2.7 X 106 8.1 X 104 a None detected in 1 ml of 1:10 dilution.

1064 HALL, BROWN, AND LEWIS APPL. MICROBIOL. total number of food specimens tested in this study (490), only 154 or 31.4% had counts above 100,000. Of these 154, 24 or 15.6% were from cheese and cheese products, 49 or 31.8% from sandwiches (most of which contained cheese), 17 or 11.0% from raw or frozen vegetables, 11 or 7.1 % from prepared dishes and convenience foods, 47 or 30.5% from raw meats, 1 or 0.6% from dry cereals and mixes, and 5 or 3.2% from frozen fish and seafood. Baby foods and frozen desserts were all below the 50,000/g level. Coliform counts. The total coliform counts, as determined by plating in VRB, ranged from none detected in 10 ml of the 1:10 dilution (less than I/g) to nearly 500,000/g. The ranges, mean, and median presumptive counts are shown in Table 2. No coliform organisms were detected in the baby foods. The highest median count (79/g) was found in raw meats, but the highest count (470,000/g) and the highest mean count (12,000/g) were from raw and frozen vegetables. The median counts indicate that over half of the specimens contained less than 100 coliform organisms per g, that over 40% (201) contained less than 10 coliform organisms per g, and that six of the nine categories had median counts of 0 (none detected) per g. In fact, the actual counts showed that 275 of the 490 specimens (56.1 %) were negative in VRB. Recommended limits for frozen precooked foods, raw meats, and some milk products are in the range of 10 or less coliform organisms per g (6, 10, 11, 13, 17; Quartermaster Food and Container Institute of the Armed Forces, MIL-M- 13966, 1955). Of the 490 specimens, 176 or 35.9% gave counts above 10/g. These were distributed among eight categories: 5 or 21.8% of 23 frozen desserts; 10 or 38.5% of 26 cheese and cheese products; 7 or 19.5% of 36 frozen fish and seafoods; 8 or 17.4% of 46 dry cereals and mixes; 23 or 46.9% of 49 raw and frozen vegetables; 9 or TABLE 2. 11.1% of 81 prepared dishes and convenience foods; 62 or 70.5% of 88 raw meats; and 51 or 47.2% of 108 sandwiches. That some of the foods contained coliform organisms at a low level that were not recovered in VRB was shown by the fact that 215 specimens yielded coliform counts in VRB, but 299 specimens were presumptively positive in LST. Of the 490 specimens tested, 215 yielded coliform counts in VRB; however, only 169 confirmed in BGLB, and some of these confirmed at lower levels than were indicated by the VRB count. The ranges and the mean and median confirmed counts are shown in Table 3. The ranges did not change appreciably; in only one instance (dry cereals and mixes) did the highest level fail to confirm. The means and median counts, however, were lower in some instances. The mean counts of cheese and cheese products, baby foods, frozen fish and sea foods, and raw and frozen vegetables remained the same, but desserts dropped from 26 to 23, dry cereals and mixes from 35 to 22, prepared dishes and convenience foods from 91 to 71, raw meats from 3,200 to 3,100, and sandwiches from 4,300 to 2,500. Seven of the nine categories had median counts of 0 (none detected) per g, whereas the two remaining categories remained the same as in the presumptive counts. Types of coliform organisms detected. The types of coliform organisms isolated from these foods were divided into four groups based on the following: their ability to produce indole from tryptophan (I); the methyl red reaction in glucose broth (M); their ability to produce acetyl methyl carbinol from glucose (V); and their ability to utilize citrate as a sole source of carbon (C). The IMViC patterns obtained related the strains to E. coli types (+ +-- or -+--), intermediate types (-+-i or ++-+), A. aerogenes types (-- +4 or - - + +), and Presumptive VRB plate counts Class of food No. of Range Mean Median specimensraemenedn Frozen desserts... 23 Oa-4. 5 X 102 26 0 Cheese and cheese products... 26 0-1.2 X 104 3.2 X 102 0 Baby foods (jars)... 33 0 0 0 Fish and seafood... 36 0-4.0 103 1.2 X 102 0 Dry cereals and mixes.46 0-6.2 X 102 35 0 Raw and frozen vegetables... 49 Prepared dishes and convenience 0-4.7 X 105 1.2 X 104 4 foods... 81 0-4.1 X 103 91 0 Raw meats... 88 0-9.5 X 104 3.2 X 103 79 Sandwiches... 108 0-4.6 X 104 4.3 X 103 9 a None detected in 10 ml of 1:10 homogenate.

VoL. 15 p1967 EXAMINATION OF MARKET FOODS FOR COLIFORMS 1065 TABLE 3. Confirmed VRB plate counts No. of Class of food seinsrange Mean Median Frozen desserts... 23 Oa 4.5 X 102 23 0 Cheese and cheese products... 26 0-1.2 X 104 3.2 X 102 0 Baby foods (jars)... 30 0 0 0 Fish and seafoods... 36 0-4.0 X 103 1.2 X 102 0 Dry cereals and mixes... 46 0-3.6 X 102b 22 0 Raw and frozen vegetables... 49 0-4.7 X 105 1.2 X 104 0 Prepared dishes and convenience foods... 81 0-4.1 X 103 71 0 Raw meats... 88 0-9.5 X 104 3.1 X 103 79 Sandwiches... 108 0-4.6 X 104 2.5 X 103 9 a None detected in 10 ml of 1:10 homogenate. b No confirmation of highest level. atypical (all bio types not falling in one of the above groups) (Table 4). In cheese and cheese products, 19 different strains were identified from 17 of the 26 specimens-12 E. coli, 3 intermediate, 1 A. aerogenes, and 1 atypical; in dry cereals and mixes, 19 were identified from 15 of the 46 specimens-no E. coli, 6 intermediate, 8 A. aerogenes, and 5 atypical; in raw meats, 78 were identified from 75 of the 88 specimens-39 E. coli (2 were enteropathogenic strains, both 026:B6), 30 intermediate, 8 A. aerogenes, and 1 atypical; in prepared and convenience foods, 29 were identified in 26 of the 81 specimens-9 E. coli (1 was an enteropathogenic strain, 0.55:B5), 13 intermediate, 4 A. aerogenes, and 3 atypical; in raw and frozen vegetables, 31 were identified in 30 of the 49 specimens-3 E. coli, 12 intermediate, 14 A. aerogenes, and 2 atypical; in sandwiches, 81 were identified in 75 of the 108 specimens-7 E. coli, 23 intermediate, 14 A. aerogenes, and 37 atypical; and in the frozen desserts, 5 were identified in' 5 of the 23 specimens-no E. coli, 1 intermediate, 4 A. aerogenes, and no atypical. No coliform organisms were isolated from the baby foods in jars. Confirmation of coliform counts and presumptive tests. Some workers accept the VRB plate count as indicative of the total coliform count without further confirmation; others feel that confirmation in BGLB Broth gives an added degree of certainty, and that confirmation in EC Medium at 44.5 C gives an acceptably accurate E. coli count (12). Similarly, confirmation of positive LST tubes in BGLB and EC is thought to indicate the actual numbers of coliform organisms and E. coli, respectively. Attempts were made in this study to determine how well the various confirmation methods functioned in accurately determining the coliform and E. coli content of the foods. A comparison of the results in VRB with con- TABLE 4. Types of coliform organisms detected in food specimens Type of food Type of coliform organisms detected Escheri- Inter- Acrobacchis coli mediate terooacr- Atypical genes Frozen desserts.. 0 1 4 0 Cheese and cheese products.. 12 3 1 3 Baby foods (jars) 0 0 0 0 Fish and seafood 3 4 4 1 Dry cereal and mixes. 0 6 8 5 Raw and frozen vegetables... 3 12 14 2 Prepared and convenience foods... 9 13 4 3 Raw meats... 39 30 8 1 Sandwiches... 7 23 14 37 Totals... 73 92 57 52 firmation in BGLB is shown in Table 5. Of the total (490) specimens tested, 215 or 43.9% gave counts in VRB but only 169 or 78.6% of these confirmed in BGLB. The percentage of confirmation ranged from 52.9% in prepared and convenience foods to 92.3% in cheese and cheese products. A comparison of the results obtained in LST with confirmation in BGLB is given in Table 6. Of the total number of specimens (490), 299 or 61% were presumptively positive in LST. This is an increase of 17.1% over the positives in VRB. Of these 299, 256 or 85.6% confirmed in BGLB, an increase of 7% as compared to confirmation in BGLB tubes inoculated from the VRB plates. The percentage of confirmation ranged from 52.9% in frozen desserts to 100% in cheese and cheese products and in sandwiches.

1066 HALL, BROWN, AND LEWIS APPL. MIICROBIOL. TABLE 5. Positive specimens in VRB and confirmed in BGLB Type of food No. of specimens Specimens positive Specimens confirmed in VRB positive in BGLB VRB_positive_in_BGLB No. Per cent' No. Per centb Frozen desserts... 23 10 43.5 6 60.0 Cheese and cheese products... 26 13 50 12 92.3 Baby foods (jars)... 33 0 0.0 Fish and seafood... 36 10 27.8 8 80.0 Dry cereals and mixes... 46 9 19.6 8 88.9 Raw and frozen vegetables... 49 26 53.1 18 69.2 Prepared and convenience foods... 81 17 21.0 9 52.9 Raw meats... 88 67 76.1 61 91.0 Sandwiches... 108 63 58.3 47 74.6 All types... 490 215 43.9 169 78.6 a Of number of specimens. Of specimens positive in VRB. TABLE 6. Positive specimens in LST and confirmed positives in BGLB Type of food No. of specimens _ Specimens positive in LST Specimens confirmed in BGLB No. Per centa No. Per cent5 Frozen desserts.... 23 17 73.9 9 52.9 Cheese and cheese products... 26 16 61.5 16 100.0 Baby foods (jars)... 33 0 0 _ Fish and seafood... 36 20 55.6 16 80.0 Dry cereals and mixes... 46 15 32.6 12 80.0 Raw and frozen vegetables... 49 35 71.4 28 80.0 Prepared and convenience foods 81 38 46.9 28 73.7 Raw meats... 88 81 92.0 70 86.4 Sandwiches... 108 77 71.3 77 100.0 All types... 490 299 61.0 256 85.6 a Of number of specimens. O of specimens positive in LST. A second comparison was made of the results obtained with LST; the number of positive tubes confirming as E. coli based on positive EC tubes was compared with the number shown culturally to contain this organism (Table 7). Of the 299 specimens positive in LST, 166 or 55.5% confirmed in EC Medium; the percentage of confirmation ranged from 20% in dry cereals and mixes to 100% in cheese and cheese products. Subcultures made from the EC Medium, however, showed a much lower confirmation as determined by IMViC results; only 73 or 43.9% of the positive EC tubes (166) confirmed as E. coli. The cultural confirmation ranged from none in dry cereals and mixes and in frozen desserts to 76.5 % in raw meats. The above findings show that coliform organisms other than E. coli gave positive EC tubes. Some of the EC tubes contained mixtures of E. coli and other coliform organisms. Such tubes are considered to have been positive because of the presence of E. coli; other positive EC tubes, however, failed to yield isolates of E. coli. The distribution of the types is shown in Table 8. Of the EC tubes, 166 were positive, but isolates were not obtained from six of these. Of the 160 from which isolates were made, 65 or 40.4% contained E. coli either alone or in combination with other organisms, 41 or 25.4% contained intermediate, 29 or 18.1 % contained A. aerogenes, and 26 or 16.1 % contained atypical strains. Of the 73 E. coli isolated from the foods, 65 or 89.0% were from positive EC tubes, but 41 or 44.5 % of the 92 intermediates, 29 or 50.9% of the 57 A. aerogenes, and 26 or 50% of the 52 atypical strains were also positive in EC broth. In this

VOL. 15, 1967 EXAMINATION OF MARKET FOODS FOR COLIFORMS 1067 TABLE 7. Positive specimens in LST and confirmed in EC broth and by culture as Escherichia coli Specimens positive Confirmed as E. coli Confirmed as E. coli Type of food No. of specimens in LST in EC by culture No. Per cent' No. Per centb No. Per centc Frozen desserts...... 23 17 73.9 6 35.3 0 0.0 Cheese and cheese products. 26 16 61.5 16 100.0 12 75.0 Baby foods (jars).3... 33.. 0.. 0.... Fish and seafood... 36 20 55.6 10 50.0 3 30.0 Dry cereals and mixes... 46 15 32.6 3 20.0 0 0.0 Raw and frozen vegetables... 49 35 71.4 19 54.3 3 15.8 Prepared and convenience foods... 81 38 46.9 18 47.4 9 50.0 Raw meats... 88 81 92.0 51 63.0 39 76.5 Sandwiches... 108 77 71.3 43 55.8 7 16.3 All types... 490 299 61.0 166 55.5 73 43.9 a of number of specimens. b Of specimens positive in LST. c Of specimens positive in EC. TABLE 8. Types of coliform organisms present in positive EC tubes Poiietubes Positive tubes Positive Positive Po.ositivE.cl with inter- tubes with tubes with No. of No.itof wthpes ol mediate A. aerogeises atypical Type of food specimens psteive wtyst types types types tubes No. Per cent No. Per cent No. Per cent No. Per cent Frozen desserts... 23 6a 0 0.0 1 33.3 2 66.7 0 0.0 Cheese and cheese products... 26 16 11 68.7 2 12.5 1 6.3 2 12.5 Baby foods (jars)... 33 0 0 0 0 0 Fish and seafood... 36 l1a 3 42.1 3 42.9 1 14.2 0 0.0 Dry cereals and mixes... 46 3 0 0.0 0 0.0 2 66.7 1 33.3 Raw and frozen vegetables... 49 19 3 15.8 7 36.8 9 47.3 0 0.0 Prepared and convenience foods... 81 18 8 44.4 7 38.8 2 11.1 1 5.5 Raw meats... 88 51 33 64.8 13 25.5 5 9.8 1 1.9 Sandwiches... 108 43 7 16.3 8 18.5 7 16.3 21 48.8 All types... 490 1668 65 40.4 41 25.4 29 18.1 26 16.1 a No isolates of coliform organisms were obtained from three of the positive EC tubes. Percentages calculated on basis of 3, 7, and 160, respectively. study, only about 40% of the positive EC tubes confirmed as E. coli; the remaining 60% were caused by other coliform organisms. Additional examinations of prewrapped sandwiches. The prewrapped sandwiches were studied as a group and were tested more extensively than the other categories. These 108 specimens were examined for the presence of coagulase-positive staphylococci, Clostridium perfringens, Bacillus cereus, salmonellae, shigellae, and enterococci. The methods used were essentially those previously described (12, 14). No isolations of salmonellae or shigellae were obtained. In 18 or 16.7% of the specimens, there were coagulase-positive staphylococci, but in numbers that were detectable by enrichment only; 10 or 9.2% contained C. perfringens, and these also were isolated by enrichment methods. Typical B. cereus strains in very low numbers were found in four or 3.7 %; counts of enterococci ranged from 220 to 450,000/g, with 37 or 34.2% containing these organisms. One-half of each sandwich was originally examined for the above organisms, and the other half was stored frozen at -15 C. Twelve of the 18 sandwiches showing the presence of coagulasepositive staphylococci were subjected to mishandling by exposing the remaining half to room temperature (24 to 27 C) for 48 hr and then testing for total, coliform, and staphylococcal counts.

1068 HALL, BROWN, AND LEWIS APPL. MICROBIOL. Total counts increased from an average of 4.2 X 106, with a range of 2 X 103 to 1.9 X 106, to an average of 2 X 108, with a range of 1.0 x 107 to 7.4 x 108. Coliform counts were lower than in the original examination in most cases, the highest level being 3.4 X 104, and the numbers of coagulase-positive staphylococci did not reach levels considered hazardous, the highest count being 5.8 X 103. DISCUSSION The results of this study of a group of foods, such as would be purchased by the housewife or other consumers, indicate marked differences in the load of viable bacteria, depending not only on the class of food but upon individual specimens within the classes. Suggested limits for total plate counts vary from 50,000 for frozen precooked foods (13) to 10,000,000 for hamburger (5), with most standards being at the level of 100,000/g (1, 6, 10, 17; Quartermaster Food and Container Institute of the Armed Forces, MIL-M-13966, 1955). Infant food formulas are the exception; total counts are not to exceed 25/ml (2). A 100,000/g standard appears to be realistic for most foods. Exceptions exist; such foods as raw vegetables, some raw meats, and raw shellfish will, by their very nature, carry more than 100,000 bacteria per g. Most of these will be saprophytic types of organisms related to the source of the food. Foods that are produced, ripened, or fermented by the action of bacteria will yield high total counts. With such foods, a known flora of nonpathogenic organisms should be present. Coliform counts as an indicator of fecal pollution must be more closely evaluated with foods than with water and milk. Many organisms not necessarily associatedwith fecal material are found on or in such foods as vegetables, fruits, and some types of meat. Suggested standards relating to the presence of coliform organisms in foods range from 0 in infant formulas (2) to 90/g in grade A dry milk (supplement 1 to Milk Ordinance and Code of 1953, Public Health Service, 1959). Frozen precooked foods may be allowed to contain none in 0.1 g (10), or 1 (6), 10 (13; Quartermaster Food and Container Institute of the Armed Forces), or 20 (17) per g depending upon the recommending agency. No limit has been suggested for foods like hamburger. The foods examined in this study, with the exception of raw meats, had coliform counts in most instances that indicated levels of 10 or less per g could be maintained. Fecal coliform counts not to exceed 230/ 100 g, as determined by an MPN technique, have been suggested as market standards for shucked oysters (National Shellfish Sanitation Work Shop, 1964). The significance of the MPN technique in which EC Medium is used is related to such factors as incubated temperature, concomitant flora, and the nature of the food. In this study, confirmation of the positive EC tubes was relatively poor, which indicated that an incubation temperature of 44.5 i 0.1 C allowed coliform organisms other than E. coli to grow out and produce gas. The fact that E. coli was isolated from any 40.4% of the positive EC tubes indicates that without confirmation an erroneously large number of food samples might be suspected of being contaminated with this fecal coliform organisms. These results are similar to those obtained by Raj and Liston (15). Incubation at a higher temperature (45.5 ±t 0.2 C) may reduce the number of false positive results, but a few false negative results may be obtained (12). It is interesting and may be of considerable significance to those examining food specimens that Geldrich et al. (7) found a closer correlation between the source of coliform and their ability to grow and produce gas at 44.5 i 0.25 C than between the type of coliform organisms and this ability. They consider the elevated temperature test to be a "fecal coliform" test rather than an E. coli-type (+ + - -) test. Opinions differ concerning the relative sensitivity of the direct plating procedure for the detection of low levels of coliform organisms and the use of liquid enrichment technique. In this study, more positive results were obtained with the LST enrichment (299 specimens) than with VRB (215). Furthermore, 256 of the positive LST tubes confirmed, as compared to 169 of the positive VRB plates. Apparently, the LST enrichment detects coliform organisms in many instances in which negative results are obtained with the VRB plating procedure. The greatest differences were noted with frozen desserts, fish and seafoods (most of which were frozen), dry cereals, and raw and frozen vegetables; this might indicate that the coliform organisms in frozen or dried products have been injured and are better able to initiate growth in a relatively noninhibitory liquid medium. Somewhat similar indications were obtained by Gunderson and Rose (8) and Wilkerson et al. (19). The observations made in this study on the use of BLGB as a liquid confirmatory medium support the almost universal acceptance of BGLB in the many standard and recommended methods for the examination of milk and milk products, food, and water and beverages (2, 12, 16). The occurrence of EEC in foods associated with outbreaks of gastroenteritis led us to examine all isolates of E. coli serologically for the 0 and B antigens most frequently associated with infantile

VOL. 15, 1967 EXAMINATION OF MARKET FOODS FOR COLIFORMS 1069 diarrhea. Two isolates of 026:B6 and one of 055:B5 were obtained. The two 026:B6 strains were obtained from cubed beef steak and pork liver, and the 055:B5, from frozen gravy and sliced beef. The organisms present in the steak and liver at levels of 410/g were the only coliform types detected, but the meat was raw, and they probably would have been killed by cooking. The gravy and sliced beef, on the other hand, was a prepared dish, and mishandling might have allowed an increase in numbers. In this instance, however, the strain was isolated only after enrichment in LST broth, the VRB plate count being negative. The results obtained indicate that the occurrence of EEC in outbreak foods is more apt to be due to contamination by a food handler than to the presence of the organisms in the food itself. Studies carried out in this laboratory showed that about 6% of 219 specimens of feces from healthy food handlers contained one or more strains of EEC (9), whereas only 0.6% of the 490 food specimens in this study contained EEC. The implication of food handlers in this type of foodborne disease is substantiated by observations in other countries (3). LITERATURE CITED 1. ABRAHAMSON, A. E., L. BUCHBINDER, J. GUENKEL, AND M. HELLER. 1959. A study of frozen precooked foods: Their sanitary quality and microbiological standards for control. Assoc. Food Drug Officials U.S. Quart. Bull. 23:63-72. 2. AMERICAN PUBLIC HEALTH ASSOCIATION. 1960. Standard methods for the examination of dairy products, p. 130, 11th ed. American Public Health Association, Inc., New York. 3. COSTIN, I. D., D. VOICULESCU, AND V. GORCEA. 1964. An outbreak of food poisoning in adults associated with Escherichia coli serotype 86:B7:H34. Pathol. Microbiol. 27:68-78. 4. EDWARDS, P. R., AND W. H. EWING. 1962. Identification of the enterobacteriaceae. Burgess Publishing Co., Minneapolis. 5. ELFORD, W. C. 1936. Bacterial limitations in ground fresh meat. Am. J. Public Health. 26:1204-1206. 6. FiTZGERALD, G. A., AND W. S. CONWAY. 1937. Sanitation and quality control in the fishery industries. Am. J. Public Health 27:1094. 7. GELDREICH, E. E., R. H. BORDNER, C. B. HUFF, H. F. CLARK, AND P. W. KABLER. 1962. Type distribution of coliform bacteria in the feces of warm-blooded animals. J. Water Pollution Control Federation. 34:295-301. 8. GUNDERSON, M. F., AND K. D. ROSE. 1948. Survival of bacteria in a precooked fresh frozen food. Food Res. 13:254-263. 9. HALL, H. E., AND G. H. HAUSER. 1966. Examination of feces from food handlers for salmonellae, shigellae, enteropathogenic Escherichia coli and Clostridium perfringens. Appl. Microbiol. 14:928-933. 10. HOBBS, B. C. 1959. Sampling and examination of foodstuffs and interpretation of results. Munici. Eng. Motor Public Health 136:469-470. 11. KELLY, C. B. 1960. Bacteriological criteria for market oysters. Robert A. Taft Sanitary Eng. Center Tech. Rept. F60-2. 12. LEWiS, K. H., AND R. ANGELoTrI. 1964. Examination of foods for enteropathogenic and indicator bacteria. U.S. Public Health Serv. Publ. 1142. 13. MASSACHUSETTS DEPARTMENT OF PUBLIC HEALTH. 1960. Rules and regulations relative to the storage and distribution of frozen foods. Bureau of Consumer Products Protection. Division of Food and Drugs. Boston. 14. MOSSEL, D. A. A., M. J. KOOPMAN, AND E. JONERIUS. 1966. The enumeration of Bacillus cereus in foods. Antonie von Leeuwenhoek J. Microbiol. Serol. 32:453-456. 15. RAJ, H., AND J. LISTON. 1961. Detection and enumeration of fecal indicator organisms in frozen sea foods. I. Escherichia coli. AppI. Microbiol. 9:171-174. 16. SHARF, J. M. 1966. Recommended methods for the microbiological examination of foods, 2nd ed. American Public Health Association, Inc., New York. 17. THATCHER, F. S. 1963. The microbiology of specific frozen foods in relation to public health: Report of an international committee. J. Appl. Bacteriol. 26:266-285. 18. UEDA, S., S. SOAKI, AND M. KOHUTO. 1959. The detection of Escherichia coli 055 in an outbreak of food poisoning. Nippon Saikingaku Zasshi 14:48-49. 19. WILKERSON, W. B., J. C. AYERS, AND A. A. KRATr. 1961. Occurrence of enterococci and coliform organisms on fresh and stored poultry. Food Technol. 15:286-292.