Coliform Species Recovered from Untreated Surface Water

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Mar. 1981, p. 657-663 0099-2240/81/030657-07$02.00/0 Vol. 41, No. 3 Coliform Species Recovered from Untreated Surface Water and Drinking Water by the Membrane Filter, Standard, and Modified Most-Probable-Number Techniquest T. M. EVANS, M. W. LECHEVALLIER, C. E. WAARVICK, AND RAMON J. SEIDLER* Department ofmicrobiology, Oregon State University, Corvallis, Oregon 97331 The species of total coliforn bacteria isolated from drinking water and untreated surface water by the membrane filter (MF), the standard most-probablenumber (S-MPN), and modified most-probable-number (M-MPN) techniques were compared. Each coliform detection technique selected for a different profile of coliform species from both types of water samples. The MF technique indicated that Citrobacter freundii was the most common coliform species in water samples. However, the fernentation tube techniques displayed selectivity towards the isolation of Escherichia coli and Klebsiella. The M-MPN technique selected for more C. freundii and Enterobacter spp. from untreated surface water samples and for more Enterobacter and Klebsiella spp. from drinking water samples than did the S-MPN technique. The lack of agreement between the number of coliforms detected in a water sample by the S-MPN, M-MPN, and MF techniques was a result of the selection for different coliform species by the various techniques. The four genera Escherichia, Klebsiella, Enterobacter, and Citrobacter are generally accepted as comprising the total coliform population (4). Enumeration of this component of the microbial aquatic ecosystem has been universally applied to document the sanitary quality of water. The usefulness of the total coliform count as an indicator of bacterial water pollution has been questioned, partly because coliform detection methods are potentially subject to interferences (14, 16). Interference with coliform detection or coliform suppression in presumptive media (11) has been thought to result from competition by noncoliform bacteria for nutrients (21). Other proposed causes of coliform suppression are the production of inhibitory products by noncoliform bacteria (16) and the failure of brilliant green lactose bile broth to recover coliforms from gas-positive presumptive tubes (3, 22). Recently, a modified most-probable-number (M-MPN) procedure was developed to document the magnitude of interference with total coliform detection in the standard MPN (S- MPN) technique (11). Coliform suppression in the presumptive and confirmed tests was found to contribute significantly to the underestimation of coliform numbers in the S-MPN technique. In addition to the quantitative impact of supt Technical paper no. 5668, Oregon Agricultural Experiment Station, Corvallis, OR 97331. pression on coliform enumeration, other factors influence the qualitative recovery of the component coliform genera. It has been illustrated with polluted specimens that the kind of water examined (sewage, unchlorinated sewage effluent, surface water), as well as media and techniques, will affect the isolation frequency of the four coliform genera (4, 9). Treatment of raw water may also influence the percentage distribution of component coliform genera found. Clark and Pagel reported that the percentage of Escherichia found in the component genera of contaminated drinking water samples was reduced compared to the untreated surface water source (4). Chlorination has been reported by others to increase the percentage of Klebsiella in the component coliform genera isolated from drinking water samples (20). The purpose of this study was to determine the influence ofstandard Methods (1) presumptive media (lactose broth, lauryl tryptose broth) and the technique used (membrane filter [MF], MPN) on the recovery of coliforn genera from chlorinated drinking water and the raw source waters. In addition, recovery of the coliform species by the new M-MPN technique was compared to those species recovered by the standard MF and MPN procedures. MATERIALS AND METHODS Sampling area. Samples were collected from the finished drinking water supply of an Oregon coastal community serving 14,000 residents and from the two 657

658 EVANS ET AL. coast range streams supplying raw water to the city. Chlorination was the only treatment applied to the raw water before entry into the distribution system (11). Chlorine was applied at a dose which resulted in an initial free chlorine concentration of approximately 1.5 ppm. Collection End microbiological techniques. Water samples were collected in sterile, 4-liter polypropylene containers with (drinking water) or without (untreated surface water) added sodium thiosulfate. Samples were placed on ice and transported back to the laboratory within 3 h after collection and analyzed within 7 h after collection. Nine- or 15-tube S-MPN analyses were performed on each sample by accepted procedures (1) using lactose broth (LB; Difco lots 652242 and 662331) and lauryl tryptose broth (LTB; Difco lot 663637) presumptive media in parallel analyses. Both presumptive media were supplemented with 18 mg of phenol red (Sigma) indicator per liter. Positive presumptive tubes were submitted to the confirmed step using brilliant green bile lactose broth (Difco lot 666632). The completed test was performed using m-endo agar LES (Difco lot 663068) and Levine eosin methylene blue (Difco lot 610498). Typical and atypical colonies were picked from the completion media and streaked onto slants containing tryptic soy broth (Difco lot 663068) supplemented with 1.5% agar (Difco) and 0.3% yeast extract (Difco lot 656810). After 24 h of incubation at 35 C, growth from the slant was removed for gram staining and transferred to secondary broth tubes containing either LB or LTB. The M-MPN technique was performed with the same media used in the S-MPN procedure with the exception that EC broth (Difco lot 641057) was added as an additional confirmatory broth and incubated at 35 C. The M-MPN scheme consisted of the S-MPN technique plus additional manipulations designed to recover coliforms from any tests which were negative in the S-MPN technique. A detailed description of the M-MPN technique can be found elsewhere (11). The MF total coliform detection technique was performed by standard procedures (1). Gelman GN-6 membranes (pore size, 0.45,um) and m-endo agar LES (Difco lot 663068) were used. Duplicate volumes of 250 and 100 ml of drinking water and 25-, 10-, and 1- ml volumes of untreated surface water were routinely analyzed. A minimum of 30% of the typical coliform colonies on MF plates were submitted to LTB for verification (2). Colonies were randomly selected for verification by beginning at the upper left quadrant of the filter. As necessary, all typical colonies from adjacent grids were picked. In some drinking water samples where the number of typical colonies was less than 10 per plate, colonies were picked from replicate plates. In this case, when possible, a total of 10 colonies were verified. Identification of total coliform bacteria. Three completed coliforms from the highest sample dilution in each presumptive medium (LTB or LB) used in S- MPN parallel analysis were selected for identification (six in total). A maximum of three coliforms recovered from false-negative S-MPN tests by the M-MPN procedure were selected for identification from the highest sample dilution in each presumptive medium (six APPL. ENVIRON. MICROBIOL. in total). In addition, 10 of the randomly selected verified total coliforms recovered by the membrane filter technique were identified. All cultures were purified by streaking onto m-endo agar LES before identification. Coliforms were identified using triple sugar iron agar slants (Difco), the IMViC (indole, methyl red, Voges-Proskauer, citrate) tests, lysine and ornithine decarboxylase broths, arginine dihydrolase broth, and malonate, rhamnose, and sorbitol fermentation (10). The biochemical reactions of the coliform isolates were compared to the reactions of coliforms obtained from the American Type Culture Collection. This comparison aided in the identification of the isolates and ensured that correct reactions occurred in the identification media. In addition, the API 20E system (Analytab Products, Plainview, N.Y.) was used to confirm the identities of 5% of the isolates. A quality assurance program as outlined in Microbiological Methods for Monitoring the Environment (2) was used throughout the course of this study. RESULTS Identification of coliform isolates. Eleven species of gram-negative, facultatively anaerobic, lactose-fermenting, and gas-producing rods were isolated by one or more of the three coliform enumeration techniques. Most of the species could readily be identified on the basis of the IMViC, lysine, ornithine, and arginine reactions (10). However, using these media, over 50% of the isolates comprising a single biotype could not be differentiated as belonging to the species Serratia liquefaciens or Citrobacter freundii. The use of additional diagnostic media (rhamnose and malonate fermentation, and deoxyribonuclease activity) indicated that these isolates were hydrogen sulfide-negative biotypes of C. freundii (Table 1). The isolates fermented malonate and rhamnose and were negative for deoxyribonuclease activity. These reactions are not typical of the lactose-fermenting biotype of S. liquefaciens (13). Data published by Davis and Ewing (6, 12) also indicate that the IMViC, lysine and ornithine decarboxylase, sorbitol, and rhamnose reactions of our isolates were representative of the hydrogen sulfide-negative biotype of C. freundii. The deoxyribonucleic acid mole percent guanine and cytosine content of the isolates designated as C. freundii was 52 to 53 mol%, which is also consistent with that of C. freundii (52.6 to 52.8 mol% [19, 24]) and not representative of S. liquefaciens (53.4 to 53.8 mol% [15]). Only 4% of the isolates identified in this study as C. freundii were hydrogen sulfide positive. Coliform organisms recovered by the MF, S-MPN, and M-MPN techniques. Over 1,300 coliform organisms (668 from untreated surface water, 668 from drinking water) were isolated

VOL. 41, 1981 and identified. The isolates were obtained from 36 untreated surface water and 193 drinking water samples collected over a 1-year period. The incidence of coliform species detected by the MF, S-MPN, and M-MPN techniques in the two types of samples examined is presented in Tables 2 and 3. The coliform species isolated by the M-MPN technique represent those obtained from completed S-MPN tests plus additional isolates recovered from false-negative S-MPN tests. The coliforn bacteria of primary interest belong to four genera: Citrobacter, Enterobacter, Escherichia, and Klebsiella (4). Cultures identified as members of these four genera comprised COLIFORM SPECIES RECOVERED FROM WATER 659 TABLE 1. Basis for identification of C. freundii 86% (M-MPN) to 95% (MF) of the total gramnegative, lactose-fennenting, gas-producing rods recovered from untreated surface water (Table 2). Other gram-negative, gas-producing rods were identified as Aeromonas hydrophilia, Hafnia alvei, S. liquefaciens, and Yersinia enterocolitica. Oxidase-positive, lactose-fermenting bacteria (Aeromonas) accounted for less than 1% of the coliform isolates. C. freundii, Escherichia coli, and Klebsiella were the most common coliforms recovered by any technique (Table 2). However, the percentage of each species in the total coliform population varied significantly with each detection technique. One would expect that selecting coliforms from the highest Coliform isolates C. freundiih2u fiheundiih2s ColHfor iolates C. H2S negativea negtive ATCCCb H2S pooitives posctfved Test or substrate 1181 Reac- % posi- Reac- % posi- Reac- % posi- Reac- % position tive tion tive e tion tive tion tive H2S (triple sugar-iron) - 0-0 + 100 + 100 Indole - 0-1.8-0 - 1.8 Methyl red + 100 + 100 + 100 + 99.4 Voges-Proskauer - 0-0 - 0-0 Citrate (Simmons) + 98 + 84.3 + 100 + 92 Lysine decarboxylase - 0-0 - 0-0 Ornithine decarboxylase + 96 + 12.5 + 69-12.5 Arginine dbiydrolase - 4 + 52-11 - 44.8 Malonate + 94-22.5 + 89-22.5 Sorbitol + 100 + 95.3 + NAf + 98.8 Rhamnose + 100 + 99.2 + NA + 99.6 a b A total of 659 isolates were tested. ATCC, American Type Culture Collection. CA total of 29 isolates were tested. 'Data taken from Davis and Ewing (6). e Data taken from Ewing and Davis (12). f NA, Data not available. TABLE: 2. Lactose-fermenting, gas-producing organisms recovered from untreated surface water samples by specific enumeration technique Membrane S-MPNa M-MPNa filter Isolate LB LTB LB LTB No. % No. % No. % No. % No. % A. hydrophilia 1 0.3 0 0 1 0.9 0 0 1 0.6 C. diversus 5 1.5 0 0 0 0 0 0 2 1.1 C. freundii 110 34.1 11 10.8 10 9.0 35 21.1 34 18.9 E. aerogenes 20 6.2 8 7.8 0 0 9 5.5 6 3.3 E. agglomerans 31 9.6 4 3.9 3 2.7 11 6.7 10 5.6 E. cloacae 2 0.6 2 2.0 7 6.3 10 6.1 10 5.6 E. coli 91 28.2 40 39.2 58 52.3 42 25.5 65 36.1 H. alvei 16 5.0 6 5.9 3 2.7 16 9.7 12 6.7 Klebsiella 45 13.9 31 30.4 27 24.3 34 20.6 31 17.2 S. liquefaciens 0 9 0 0 0 0 1 0.6 1 0.6 Y. enterocolitica 2 0.6 0 0 2 1.8 7 4.2 8 4.4 a Parallel MPN analyses conducted on each sample using LB and LTB presumptive media.

660 EVANS ET AL. APPL. ENVIRON. MICROBIOL. TABLE 3. Lactose-fermenting, gas-producing organisms recovered from drinking water samples by specific enumeration techniques S-MPNa M-MPNa Membrane Isolate filter LB LTB LB LTB No. % No. % No. % No. % No. % A. hydrophilia 0 0 1 2.5 2 4.9 2 1.3 2 1.5 C. diversus 1 0.3 0 0 0 0 0 0 2 1.5 C. freundii 298 79.3 30 75.0 25 61.0 97 61.4 85 63.4 E. aerogenes 6 1.6 1 2.5 0 0 4 2.5 6 4.5 E. agglomerans 15 4.0 1 2.5 0 0 9 5.7 3 2.2 E. cloacae 17 4.5 2 5.0 1 2.4 21 13.3 4 3.0 E. coli 16 4.3 3 7.5 4 9.8 5 3.2 14 10.4 H. alvei 8 2.1 0 0 6 14.6 5 3.2 6 4.5 Klebsiella 9 2.4 2 5.0 3 7.3 14 8.9 12 9.0 S. liquefaciens 2 0.5 0 0 0 0 0 0 0 0 Y. enterocolitica 4 1.1 0 0 0 0 1 0.6 0 0 a Parallel MPN analyses conducted on each sample using LB and LTB presumptive media. sample dilutions in the fermentation tube procedures would result in the predominating coliform organisms being identified. The results show that indeed this did not happen. That is, the most dominant organisms in the original samples should have been those most often detected by the MF technique. C. freundii comprised the greatest percentage of the coliforms recovered by the MF technique, whereas E. coli and Klebsiella dominated the isolates detected by the two fermentation tube techniques. Due to its increased sensitivity in detecting coliforms, the M-MPN technique recovered a significantly different profile of coliform species than the S-MPN technique. The difference was reflected in the greater frequency of Citrobacter, Enterobacter, Hafnia, and lactose-fermenting Y. enterocolitica isolation by the M-MPN technique. Both techniques recovered comparable numbers of E. coli and Klebsiella (Table 2). The presumptive broths (LTB, LB) used in the fermentation tube procedures also influenced the incidence of coliform species recovered from untreated surface water samples. For example, in the S-MPN technique, E. coli was recovered at a greater frequency with LTB whereas Klebsiella and Enterobacter aerogenes were recovered more often with LB presumptive media (Table 2). In the M-MPN technique, LTB had a similar selective advantage over LB in the isolation of E. coli. However, other coliform species were recovered at comparable rates with both presumptive media. Escherichia, Klebsiella, Citrobacter, and Enterobacter species comprised 95 to 98% of the lactose-fermenting, gas-producing, gram-negative bacteria recovered from drinking water by the three coliformn detection techniques (Table 3). The use of LTB presumptive medium in the S-MPN resulted in only 80% of the coliform isolates being classified as these four genera due to the detection of H. alvei. Lactose-fermenting, gas-producing isolates which did not belong to the four genera of accepted coliforms were identified as A. hydrophilia, H. alvei, S. liquefaciens, and Y. enterocolitica. None of the verified colonies recovered by the MF technique were identified as A. hydrophilia. A. hydrophilia did account for between 1.3 and 4.9% of the gas-producing isolates recovered by the MPN techniques. In drinking water, C. freundii was clearly the most dominant coliform species recovered by any of the techniques (Table 3). However, the profile of coliform species detected differed for each enumeration technique. As observed for surface water, E. coli and Klebsiella each comprised a greater percentage of the isolates detected by the fermentation tube procedures than by the MF technique in drinking water samples (Table 3). The percentages of Enterobacter spp. and Klebsiella in the coliform population were greater with the M-MPN procedure than with the S-MPN technique. E. coli and C. freundii each represented comparable percentages of the coliform species detected by the two fermentation tube techniques. Except for minor differences with H. alvei and C. freundii, the profile of coliform species recovered by the S-MPN technique from drinking water was not significantly different for the two presumptive media (LTB, LB) used (Table 3). On the other hand, the use of LB in the M-MPN technique was distinctly inferior to LTB in the recovery of E. coli. LTB was inferior to LB in the M-MPN for recovery of Enterobacter agglomerans and Enterobacter cloacae. Differences in numbers of each coliform spe-

VOL. 41, 1981 cies detected by the M-MPN and S-MPN techniques are reflected by the incidence of the species isolated from false-negative S-MPN tests. Citrobacter, Enterobacter, Hafnia, Yersinia, Escherichia, Klebsiella, Serratia, and Aeromonas, in decreasing order of occurrence, have been isolated from false-negative S-MPN tests of untreated surface water (Table 2). Citrobacter, Enterobacter, Klebsiella, Escherichia, Hafnia, and Yersinia, in decreasing order of occurrence, were isolated from false-negative S- MPN tests of drinking water (Table 3). Different presumptive media also influenced the coliform species recovered from false-negative S-MPN tests. For example, with surface water, E. coli and E. aerogenes were isolated more frequently from gas-negative LTB presumptive tubes than from gas-negative LB tubes. The number of C. freundii, E. agglomerans, H. alvei, Klebsiella, S. liquefaciens, and Y. enterocolitica isolations from false-negative S- MPN tests was comparable with either presumptive medium. In drinking water samples, E. aerogenes and E. coli were isolated more frequently from false-negative S-MPN tests when LTB presumptive medium was used and E. agglomerans, E. cloacae, and H. alvei when LB presumptive medium was used. C. freundii and Klebsiella were isolated at comparable rates with either presumptive medium. DISCUSSION The identification of coliforms isolated from high-quality surface water and drinking water by the MF, S-MPN, and M-MPN techniques demonstrated that each technique selected for a specific group of coliforms. A comparison of the coliform species identified from the same samples indicated that the most numerous species recovered by the MF technique was C. freundii, whereas the two fermentation tube techniques displayed a selectivity for E. coli and Klebsiella. These results differed from those reported by Dutka and Tobin (9) who found that the MF technique relative to the S-MPN procedure selected for E. coli and Klebsiella spp. in creek water. Several factors may explain why the MF and the two fermentation techniques in this study recovered a different profile of coliform species from that reported by Dutka and Tobin. The degree of pollution in the stream (geometric mean of 28,000 total coliforms per 100 ml) studied by Dutka and Tobin was greater than in the stream (geometric mean of 160 total coliforms per 100 ml) examined in this investigation. In addition, Dutka and Tobin reported that Citrobacter was not among the coliform genera isolated from creek water by the MF COLIFORM SPECIES RECOVERED FROM WATER 661 technique. Nevertheless, the results of this and other investigations (9, 23) support the conclusion that the type of water examined and the type of pollution in that water influences the selectivity patterns of the various coliform detection techniques. A significant difference was found in the profiles of coliform species recovered from drinking water and from the untreated surface water source. For example, C. freundii, E. coli, and Klebsiella were the dominant coliform -species recovered from untreated surface water, whereas C. freundii alone was the dominant species recovered from contaminated drinking water samples. All coliform detection techniques reflected this change in the profile of coliforin species obtained from the two types of water supplies. The presence of high numbers of C. freundii in contaminated drinking water samples may reflect a resistance to chlorine by this organism relative to other coliforms. High numbers of the hydrogen sulfide-negative biotype of C. freundii have been reported in other aquatic environments exposed to chlorine (7). The observed differences in the profiles of coliforn species obtained by the MF and S- MPN techniques were more pronounced for surface water than for drinking water samples. These observations are consistent with coliforn enumeration studies where the geometric mean coliform numbers recovered by the MF and S- MPN were the same for drinking water, but for surface water the MF geometric mean was nearly threefold higher than the S-MPN (11). The primary reason for the difference in geometric mean numbers for surface waters was the increased detection of Citrobacter by the MF technique. The more pronounced differences in the profiles of coliforms obtained by the M- MPN and S-MPN were caused by failures in the S-MPN to recover as many coliforn species in some 85 to 90% of the drinking water and surface water samples (11). Coliforms most frequently recovered from false-negative S-MPN tests and those which were responsible for the differences in species profiles were Citrobacter, Enterobacter, and Klebsiella. The increased percent recovery of these genera should not obscure the fact that the M-MPN also outperforned the S-MPN in the actual number of E. coli isolates detected. In addition, the number of detectable coliformpositive drinking water samples based on the S- MPN or MF techniques was doubled by the M- MPN (11). This doubling in incidence and the fivefold increase in the M-MPN geometric mean over standard techniques was a result of the increased detection of the three coliform genera. A 9- or 15-tube fermentation tube technique

662 EVANS ET AL. was used in the present study whereas drinking water-monitoring laboratories have the option of using a 5-tube procedure. The majority of the false-negative S-MPN tests observed with contaminated drfinking water samples occurred in the fermentation tubes inoculated with 10-ml portions. Therefore, the selectivity patterns observed in the S-MPN in this investigation would occur in monitoring laboratories as well. The nature of the enrichment processes occurring in the two fermentation tube techniques may explain why the S-MPN and M-MPN procedures recovered different groups of coliform species. Selection against a certain coliform species in the fermentation tube procedure may result from failure to produce gas in presumptive media due to unfavorable nutritional conditions (3, 5), competition for lactose (21), or the production of inhibitory products by noncoliform bacteria (16), and from the inhibitory nature of the confirmatory medium (22), or the failure to produce typical colonies on the agar medium used in the completed step (11). The use of different presumptive media in the fermentation tube techniques and the use of m- Endo agar LES in the MF technique also influenced the selectivity patterns. LB, LTB, and m- Endo agar LES all yielded different profiles of coliform species. Other investigators have also reported that the media used in coliform detection techniques greatly influenced the profile of coliforms obtained (4, 8, 9). Data presented here and elsewhere (11) indicated that the selection for or against certain coliformn species, i.e., Citrobacter, Enterobacter, and Klebsiella species, influenced the number and species of coliforms detected in contaminated drinking water samples and influenced compliance with the Safe Drinking Water Act. The use of different coliform detection methodologies will result in different numbers of coliforms being recovered from aquatic environments (raw sewage, sewage effluent, creeks, and chlorinated drinking water), where different coliforn species predominate (17). Additional studies must be conducted to determine if the selection for or against certain coliform species in the S-MPN technique and the resulting false-negative results are commonly occurring phenomena. Such investigations would underscore the necessity for revising current MF and MPN coliform enumeration techniques. The failure of the S-MPN technique to detect Enterobacter and Citrobacter species in chlorinated drinking water is disturbing and could result in such supplies being classified as potable. The occurrence of any coliforn species in drinking water should always cause concern. Citrobacter and Enterobacter are present at 105 to APPL. ENVIRON. MICROBIOL. 106 cells per gram of human feces (18) and are therefore also present in sewage. Their presence is at least a signal of inadequate raw water treatment or perhaps is a result ofcontamination from within the distribution system. Therefore, any selection against these or any other coliforms by a coliform detection technique could have public health significance. The possibility of selection for or against certain coliform groups should not be overlooked when new and more sensitive coliform detection techniques are being developed. ACKNOWLEDGMENTS This research was supported by cooperative agreement CR 806287 from the U.S. Environmental Protection Agency, Drinking Water Research Division. The suggestions and contributions of Harry D. Nash, U.S. Environmental Protection Agency, Office of Drinking Water Research, were greatly appreciated. LITERATURE CITED 1. American Public Health Association. 1975. Standard methods for the examination of water and waste water, 14th ed. American Public Health Association, Inc., Washington, D.C. 2. Bordner, R., and J. Winter (ed.). 1978. Microbiological methods for monitoring the environment. U.S. Environmental Protection Agency, Cincinnati. 3. Chambers, C. W. 1950. Relationship of coliform bacteria to gas production in media containing lactose. Public Health Rep. 65:619-627. 4. Clark, J. A., and J. E. Pagel. 1977. Pollution indicator bacteria associated with municipal raw and drinking water supplies. Can. J. Microbiol. 23:465-470. 5. Darby, C. W., and W. L Mallmann. 1939. Studies on media for coliform organisms. J. Am. Water Works Assoc. 31:689-706. 6. Davis, B. R., and W. H. Ewing. 1971. Biochemical characterization of Citrobacter freundii and Citrobacter diversus. DHEW Publication No. (CDC)76-8283. U.S. Department of Health, Education, and Welfare, Atlanta. 7. Davis, E. M., and S. R. Keen. 1974. Municipal wastewater bacteria capable of surviving chlorination. Health Lab. Sci. 111:268-274. 8. Dutka, B. J. 1973. Coliforms are an inadequate index of water quality. J. Environ. Health 36:39-48. 9. Dutka, B. J., and S. E. Tobin. 1976. Study on the efficiency of four procedures for enumerating coliforms in water. Can. J. Microbiol. 22:630-635. 10. Edwards, P. R., and W. H. Ewing. 1972. Identification of Enterobacteriaceae. Burgess Publishing Co., Minneapolis. 11. Evans, T. M., C. E. Waarvick, R. J. Seidler, and M. W. LeChevallier. 1981. Failure of the most-probablenumber technique to detect coliforms in drinking water and raw water supplies. Appl. Environ. Microbiol. 41: 130-138. 12. Ewing, W. H., and B. R. Davis. 1972. Biochemical characterization of Citrobacter diversus (Barkey/ Werkman and Gillen) and designation of the neotype strain. Int. J. Syst. Bacteriol. 22:12-18. 13. Ewing, W. H., B. RI Davis, M. A. Fife, and E. F. Lessel. 1973. Biochemical characterization of Serratia liquefaciens (Grimes and Hennerty) Bascomb et al. (formerly Enterobacter liquefaciens) and Serratia rubidaea (Stapp) comb. nov. and designation of type and neotype strains. Int. J. Syst. Bacteriol. 23:217-225.

VOL. 41, 1981 14. Geldreich, E. E., M. J. Allen, and R. H. Taylor. 1978. Interferences to coliform detection in potable water supplies, p. 13-20. In C. W. Hendricks (ed.), Evaluation of the microbiology standards for drinking water. U. S. Environmental Protection Agency, Washington, D.C. 15. Grimont, P. A. D., F. Grimont, and M. P. Starr. 1978. Serratia proteamaculans (Paine and Stansfield) comb. nov., a senior subjective synonym of Serratia liquefaciens (Grimes and Hennerty) Bascomb et al. Int. J. Sys. Bacteriol. 28:503-510. 16. Hutchison, D., R. H. Weaver, and M. Scherago. 1943. The incidence and significance of microorganisms antagonistic to Escherichia coli. J. Bacteriol. 45:29. 17. Kabler, P. 1954. Water examinations by the membrane filter and most probable number procedures. Am. J. Public Health 44:379-386. 18. Leclerc, H., D. A. Mossel, P. A. Trinel, and F. Gavini. 1976. Microbiological monitoring-a new test for fecal contamination, p. 23. In A. W. Hoadley and B. J. Dutka (ed.), Bacterial indicators/health hazards associated COLIFORM SPECIES RECOVERED FROM WATER 663 with water. American Society for Testing and Materials, Philadelphia. 19. Popoff, M., and G. H. Stoleru. 1980. Taxonomic study of Citrobacter freundii biochemical variants. Ann. Microbiol. 131:189-218. 20. Ptak, D. J., W. Ginsburg, and B. F. Willey. 1973. Identification and incidence of Klebsiella in chlorinated water supplies. J. Am. Water Works Assoc. 65:604-608. 21. Reitler, R., and R. Seligmann. 1957. Pseudononas in drinking water. J. Appl. Bacteriol. 20:145-150. 22. Ruchhoft, C. C. 1935. Comparative studies of the media for determination of the coli-aerogenes group in water analysis. J. Am. Water Works Assoc. 27:1732-1745. 23. Shipe, E. L., and G. M. Cameron. 1954. A comparison of the membrane filter with the most probable number method for coliform determinations from several waters. Appl. Microbiol. 2:85-88. 24. Starr, M. P., and M. Mandel. 1969. DNA base composition and taxonomy of phytopathogenic and other enterobacteria. J. Gen. Microbiol. 56:113-123. Downloaded from http://aem.asm.org/ on January 10, 2019 by guest