APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 1979, p. 351-3580099-2240/79/09-0351/08$02.00

Vol. 38, No. 3

Comparison of m-Endo LES, MacConkey, and Teepol Mediafor Membrane Filtration Counting of Total Coliform Bacteria

in WaterW. 0. K. GRABOW* AND MARTELLA DU PREEZ

National Institute for Water Research, Council for Scientific and Industrial Research,Pretoria 0001, South Africa

Received for publication 2 April 1979

Total coliforn counts obtained by means of standard membrane filtrationtechniques, using MacConkey agar, m-Endo LES agar, Teepol agar, and padssaturated with Teepol broth as growth media, were compared. Various combi-nations of these media were used in tests on 490 samples of river water and citywastewater after different stages of conventional purification and reclamationprocesses including lime treatment, sand filtration, active carbon treatment,ozonation, and chlorination. Endo agar yielded the highest average counts for allthese samples. Teepol agar generally had higher counts then Teepol broth,whereas MacConkey agar had the lowest average counts. Identification of 871positive isolates showed that Aeromonas hydrophila was the species most com-monly detected. Species of Escherichia, Citrobacter, Klebsiella, and Enterobac-ter represented 55% of isolates which conformed to the definition of total coliformson Endo agar, 54% on Teepol agar, and 45% on MacConkey agar. Selection forspecies on the media differed considerably. Evaluation of these data and literatureon alternative tests, including most probable number methods, indicated that thetechnique of choice for routine analysis of total coliform bacteria in drinkingwater is membrane filtration using m-Endo LES agar as growth medium withoutenrichment procedures or a cytochrome oxidase restriction.

The total coliform count is one of the mostuseful indicators of water pollution (2, 9, 19, 28,33, 44, 47). Although various members of thecoliform group of bacteria may multiply in en-vironments other than the gastrointestinal tract,they are excreted in large numbers in the fecesof warm-blooded animals and their presence inwater is associated with fecal pollution (4, 19).An important feature of coliform bacteria is thatthey are detectable by relatively simple, rapid,and cheap techniques. Unfortunately, however,a wide variety of techniques which differ inaccuracy and reliability are being used to eval-uate coliforms in water (5, 22, 34). Consequently,results from different laboratories can hardly becompared, the implementation of water qualitystandards is of limited value, and unreliabletechniques may even have far-reaching healthinplications (5, 6, 22).

In this study, counts obtained by means ofthree growth media commonly used in standardmembrane filtration (MF) tests for total coli-forms in water have been compared. The resultscontribute to information urgently needed forthe standardization of coliform techniques (5, 6,16, 22). The media involved were m-Endo LES

agar, used in countries such as the United States(2), Canada (33), and West Germany (3), Teepolmedia, used in Britain (9), and MacConkey agar,used in South Africa (22, 44) and Canada (33).

MATERIALS AND METHODSMF techniques. Laboratory procedures complied

with the specifications of the American Public HealthAssociation (2), Department of Health and Social Se-curity, London, England (9), and the South AfricanBureau of Standards (44). Sartorius filter holders,Gelman GN-6 membranes (pore size, 0.45 um), anddisposable plastic petri dishes (diameter, 65 mm) withloose-fitting lids were used. All counts are expressedas the average of tests done in triplicate. Incubationwas in a circulating air incubator at 35 + 0.5°C for 20to 24 h.

Media. m-Endo agar LES (code 0736-01; Difco Lab-oratories) was obtained commercially. Membrane-en-riched Teepol broth (0.4 ET) was prepared as specifiedby the Department of Health and Social Security (9)and contained 40 g of peptone (Difco), 6 g of yeastextract (Difco), 30 g of lactose (Merck & Co., Inc.), 50ml of a 0.4% aqueous solution of phenol red, 4 ml ofTeepol 610 (British Drug Houses), and 1,000 ml ofdistilled water. Membranes were either incubated onpads saturated with this medium or on the mediumsolidified with 1.5% agar. MacConkey agar was pre-

351

on April 20, 2019 by guest

http://aem.asm

.org/D

ownloaded from

352 GRABOW AND DU PREEZ

pared by the following formula (26): 15 g of agar(Difco), 20 g of peptone (Difco), 10 g of lactose(Merck), 5 g of bile salts (Difco no. 3), 5 g of sodiumchloride, 0.12 g of bromocresol purple, and 1,000 ml ofdistilled water.Water samples analyzed. Samples of settled

wastewater, biofilter effluent, activated sludge ef-fluent, and a mixture of biofilter and activated sludgeeffluents before and after chlorination to a total chlo-rine content of about 4 mg/liter, followed by sandfiltration, were collected at the Daspoort wastewaterpurification works in Pretoria, South Africa (29). TheApies River was sampled shortly before and after thedischarge of secondary treated wastewater (humustank effluent, unchlorinated) from the Daspoort puri-fication works (27). Samples were also collected fromthe following stages in an experimental 4,500 m3/daymultiple-barrier wastewater reclamation plant (Stan-der plant) in Pretoria (25): influent (secondary treatedwastewater) (DR1); after lime treatment at pH levelsranging from 9.6 to 11.4 (DR3); after quality equali-zation in a pond with mean residence time of about 10h (DR5); after sand filtration (DR11); after breakpointchlorination (DR12); after active carbon treatment(DR13); after final chlorination (DR14); and afterozonation (DR15). The latter was at times used fordisinfection instead ofbreakpoint chlorination. Bottlesused for collecting samples of chlorinated water con-tained sodium thiosulfate for dechlorination (2). Sam-ples were collected during the period from December1976 to January 1979. They were homogenized in aSanyo mixer for 4 min at a speed selector setting of 4(26) and processed within 3 h after collection.

Identification of coliform-like bacteria. Mem-branes with 20 to 50 well-spaced coliform-positivecolonies were selected. All coliform-positive colonieswere picked from these membranes and purified onthe same medium for identification by means of thecommercial API 20E system (30, 40). The IMVic (in-dole, methyl red, Voges-Proskauer, citrate) (2) andcytochrome oxidase (15) tests were done additionallyon all these isolates. Names of isolates in Table 3which do not appear in the 8th edition of Bergey'sManual of Determinative Bacteriology are used bythe manufacturers of the API test system (AnalytabProducts Inc., Plainview, N.Y.).

RESULTSColiform counts on wastewater and river

water. Average total coliform counts obtainedon MacConkey, Teepol, and Endo agars for sam-ples collected at the Daspoort wastewater puri-fication works are listed in Table 1. MacConkeyand Endo agars were compared in tests on 70 ofthese samples. Endo agar yielded the highestaverage count for each sampling point (Table 1).The highest count for 51 (73%) of the 70 sampleswas recorded on Endo agar, whereas Mac-Conkey agar yielded the highest count for 19(27%) of the samples. MacConkey, Teepol, andEndo agars were compared in tests on 22 sam-ples of secondary effluent before and after chlo-

rination and of the river upstream and down-stream of the discharge from the purificationworks (Table 1). Endo agar yielded the highestaverage count for each of the sampling stations,followed by Teepol and MacConkey agars.Endo, Teepol, and MacConkey agars had thehighest counts for 17 (77%), 3 (14%), and 2 (9%)of the 22 samples, respectively. The differencein counts on different media was relatively con-stant for all sampling points, which indicatesthat the quality ofwater had no significant effecton the ability of the media to support the pro-duction of colonies which conform to the defi-nition of total coliforms.Coliform counts on samples from the

Stander wastewater reclamation plant. Allthe media concerned were compared in variouscombinations in total coliform tests on samplesfrom different treatment stages of the Standerwastewater reclamation plant. Table 2 showsthat, apart from a few minor exceptions, thehighest average count was recorded on Endoagar, followed by Teepol agar, Teepol broth, andMacConkey agar. Endo and MacConkey agarswere compared in tests on 90 samples (Table 2).Endo agar yielded the highest count for 56 (62%)of the 90 samples, and MacConkey agar yieldedthe highest count for 34 (38%). However, Endoagar had 0 counts for 12 of the samples, andMacConkey agar had 0 counts for only 4.MacConkey agar, Teepol broth, and Teepol agarwere compared in tests on 96 samples of theStander plant (Table 2). All three media had 0counts for six of the samples. Teepol agar, Tee-pol broth, and MacConkey agar had the highestcounts for 49 (55%), 38 (42%), and 3 (3%) of theremaining 90 samples, respectively. Teepol agaryielded 0 counts for a total of five samples, andMacConkey agar and Teepol broth each forseven samples. All four media were compared intests on another 53 samples ofthe Stander plant.One sample had 0 counts on all media. Endoagar, Teepol agar, Teepol broth, and Mac-Conkey agar had the highest counts for 48 (92%),4 (8%), none (0%), and none (0%) of the remain-ing 52 samples, respectively.The large differences in coliform counts for

various series of tests on DR3 (Table 2) are dueto variations in lime dosage which were intro-duced for experimental purposes at the timewhen these counts were done (25). The limedosage also affected coliform counts in the nexttwo treatment units, namely, the quality equal-ization pond (DR5) in which a relatively highpH level was maintained for about 10 h (25) andthe sand filters (DR1l). Samples for each seriesof comparative coliform tests on DR3, DR5, andDR1l were taken when lime dosage was at a

APPL. ENVIRON. MICROBIOL.

on April 20, 2019 by guest

http://aem.asm

.org/D

ownloaded from

TECHNIQUES FOR TOTAL COLIFORMS 353

TABLE 1. Total coliform counts for wastewater and river water obtained by means ofmembrane filtrationusing different growth media

No. of Total coliform count/100 mlaSource of samplesapesamples MacConkey agar Teepol agar m-Endo LES agar

Settled wastewater 13 193 x 1o5; (50 to 340) 253 x 105; (90 to 450) xX 105 105

Biofilter effluent 7 149 x 104; (350 to 520) 178 x 104; (36 to 730) xX 104 104

Activated sludge 7 184 x 103; (80 to 360) 209 x 103; (110 to 390)effluent x 103 X 103

Effluent before 9 148 x 103; (24 to 500) 200 x 103; (70 to 700) xchlorinationb X 103 103

5 734 x 103; (250 to 852 x 103; (270 to 918 x 103; (390 to 1,800)1,900) X 103 2,200) x 103 X 103

Effluent after 11 1,173; 0-6,500 2,316; 0-9,800chlorinationc

5 778; 190-2,000 1,710; (250 to 5,000) 2,286; 630-6,000River before 11 31 x 103; (6 to 83) x 54 x 103; (5 to 190) x

discharged 103 1036 32 x 103; (7 to 60) x 92 x 103; (15 to 213) x 129 x 103; (12 to 360) x

103 103 103River after discharge' 12 458 x 103; (51 to 983 x 103; (25 to 5,400)

2,000) x 103 X 1036 114 x 103; (8 to 400) x 208 x 103; (20 to 600) 342 x 103; (33 to 1,100)

i03 X103 X 103a Each value indicates average, followed by range of counts.b Mixture of Daspoort biofilter and activated sludge effluents.Mixture described in footnote b after chlorination and sand filtration.

d. e Apies River upstream and downstream of secondary treated discharge from Daspoort purification works.

TABLE 2. Total coliform counts for samples from various stages in a wastewater reclamation plant,obtained by means ofmembrane filtration with various growth media

No. of Total coliform count/100 mlaSource of sample sapesamples MacConkey agar Teepol broth Teepol agar m-Endo LES agar

Intake (DR1)b 32 172 x 103; (13 to 191 x 103; (16 to800) x 103 1,020) x 103

15 137 x 103; (17 to 213 x 103; (43 to 243 x 103; (65 to320) x 103 450) x 103 440) x 103

10 220 x 103; (50 to 393 x 103; (170 to 385 x 103; (180 to 618 x 103; (300 to590) x 103 760) x 103 590) x 103 1,280) x 103

After lime 17 432; 0-3,500 864; 0 to 3,800treatment (DR3) 14 3,077; 33-19,000 3,296; 60-17,000 4,276; 20-28,000

7 62 x 103; (22 to 127 x 103; (35 to 149 x 103; (79 to 215 x 103; (130 to100) X 103 200) x 103 200) x 103 350) x 103

After quality 17 214; 0-1,400 517; 0-3,100equalization 16 2,595; 0-17,200 2,760; 0-14,500 3,069; 0-20,800(DR5) 8 42 x 103; (24 to 72 x 103; (45 to 73 x 103; (15 to 134 x 103; (72 to

88) x 103 116) x 103 135) x 103 190) X 103After sand filtration 24 364; 0-1,300 730; 0-8,200(DR1l) 16 2,309; 0-9,200 2,743; 0-10,200 2,976; 0-19,500

6 9.8 x 103; (0.8 to 17.3 x 103; (5.6 to 21.9 x 103; (6.9 to 34.7 x 103; (19.014.3) x 103 20.0) x 103 51.0) x 103 to 76.0) x 103

After ozonation 7 4; 0-13 4; 0-24 4; 0-14(DR15) 8 6; 0-23 5; 0-26 5; 0-24 27; 0-92

After carbon 28 41; 0-260 57; 0-460 69; 0-440treatment 14 61; 1-360 92; 1-460 93; 1-320 260; 3-620(DR13)a Each value indicates average, followed by range of counts.b The raw water intake was activated sludge effluent.

VOL. 38, 1979

on April 20, 2019 by guest

http://aem.asm

.org/D

ownloaded from

354 GRABOW AND DU PREEZ

constant level. The seven samples of DR3 inwhich all four media were compared (Table 2)were taken during a period of low-lime dosage(operational pH in lime treatment unit about9.6), whereas the 17 samples of DR3 in whichMacConkey and Endo agars were comparedwere taken during a period of high-lime dosage(operational pH level in lime treatment unitabout 11.2).

In comparative tests on 49 samples of watertaken after intermediate chlorination (DR12)and final chlorination (DR14), MacConkey andEndo agars yielded positive results for five andtwo samples, respectively. In tests on another 24samples of chlorinated water, Teepol agar,MacConkey agar, and Teepol broth yielded pos-itive results for four samples, two samples, andone sample, respectively. All four media werecompared in tests on 14 other samples of chlo-rinated water, but none of them yielded positiveresults. The general quality of the chlorinatedwater compared favorably with that of conven-tional drinking water supplies, and coliformswere recorded on occasions when chlorinationwas intentionally inefficient for research pur-poses (23). The same applied to ozonation (23).The results of tests on chlorinated and ozonatedwater indicate that Teepol and MacConkeyagars tended to yield positive results more oftenthan did Endo agar. However, the differencebetween positive and negative results was gen-erally one or a few colonies, and the number ofsamples which yielded these differences werenot enough to warrant statistically significantconclusions. Apart from the uncertainty aboutchlorinated and ozonated water, the results oftests on the other samples from the Standerplant showed that the treatment of the waterhad no significant effect on the relative differ-ences in coliform counts on the different media(Table 2).The following data on the highest counts for

individual samples irrespective of source werederived from the above-mentioned tests, as wellas additional tests which were excluded for sta-tistical reasons from Tables 1 and 2 because theywere done on a series of tests comprising fewerthan five samples. Teepol broth had highercounts than MacConkey agar for 131 samples,whereas MacConkey agar had higher countsthan Teepol broth for 30 samples. Teepol agarhad higher counts than Teepol broth for 93samples, whereas Teepol broth had highercounts than Teepol agar for 59 samples. Endoagar had higher counts than Teepol agar for 81smples, whereas Teepol agar had higher countsthan Endo agar for 12 samples. Endo agar hadhigher counts than MacConkey agar for 179

samples, whereas MacConkey agar had highercounts than Endo agar for 54 samples.

Identification of coliform-like isolates.Isolates recorded as coliform positive were iden-tified in tests on 13 samples of water in whichMacConkey, Teepol, and Endo agars were usedas growth media. The samples comprised two ofDaspoort secondary-treated wastewater, threeof the Apies River before the Daspoort dis-charge, one ofthe Apies River after the Daspoortdischarge, two of the Daspoort secondarytreated wastewater after chlorination and sandfiltration, one of the raw water intake (DR1),three of the effluent of the lime treatment unit(DR3), and one of the effluent of the ozonationunit (DR15) in the Stander wastewater recla-mation plant. The identities of the 871 isolatesare combined in Table 3, since there was nosignificant difference in the relative numbers ofspecies isolated from samples of differentsources. Among the 225 isolates from Mac-Conkey agar, 116 (45%) were true coliforms(Escherichia coli and species of Klebsiella, En-terobacter, and Citrobacter), whereas 135 (53%)were Aeromonas hydrophila and the other 4(2%) were various other species. The 275 isolatesfrom Teepol agar consisted of 149 (54%) truecoliforms, 114 (41%) A. hydrophila, and 12 (4%)other species. The 341 coliform-like coloniespicked from membranes incubated on Endo agarcomprised 186 (55%) true coliforms, 138 (40%)A. hydrophila, and 17 (5%) other species.General features of the media. Colonies

which conformed to the definition of coliformswere generally easy to recognize on all the me-dia. The golden-green metallic sheen of the col-onies on Endo agar was occasionally difficult toidentify. Differentiation of yellow coliform colo-nies from pink and light red colonies on Teepolmedia was at times uncertain, particularly oncrowded membranes. Using pads saturated withTeepol broth proved time consuming, tediousand inconvenient. The pads cannot be preparedin advance and stored for immediate use like theagar-based media. The most important disad-vantage of the saturated pads was that theytended to dry out during incubation, which ofteninterfered with counts. One advantage of Endoagar is that the medium does not require steri-lization (2). MacConkey agar has to be auto-claved for 15 min (44), and Teepol broth has tobe steamed for 30 min on 3 successive days (9).On the other hand, Endo agar is a complexmedium, and it is advisable to only use goodquality media from reliable commercial sup-pliers (16). MacConkey and Teepol media areeasily prepared from basic ingredients in thelaboratory. Teepol medium was cheaper than

APPL. ENVIRON. MICROBIOL.

on April 20, 2019 by guest

http://aem.asm

.org/D

ownloaded from

TECHNIQUES FOR TOTAL COLIFORMS 355

TABLE 3. Identity ofpositive isolates in total coliform tests on 13 water samples of various origins, usingmembrane filtration and three different growth media

Positive isolates on the following medium:

Isolate MacConkey agar Teepol agar Endo agar

No. %a No. %a No. %a

E. coli 43 17 48 17 44 13C. freundii 24 9 23 8 28 8C. diversus 2 <1Citrobacter Spp.b 1 <1 1 <1K. pneumoniae 20 8 29 11 33 10K. ozaenae 3 1 1 <1K. rhinoscleromatis 2 <1Klebsiella spp.b 1 <1E. aerogenes 2 <1 2 <1E. cloacae 9 4 25 9 52 15Enterobacter spp.b 1 <1Erwinia herbicola (Enterobacter 14 5 22 8 21 6

agglomerans)

Total for the above typical coliforms 116 45 149 54 186 55

Y. enterocolitica 1 <1 1 <1Shigella dysenteriae 1 <1Shigella Spp.b 1 <1Serratia liquefaciens 6 2 3 <1S. rubidaea 1 <1 1 <1S. marcescens 3 <1Alcaligenes faecalis 2 <1Alcaligenes Spp.b 4 1Flavobacterium meningosepticum 2 <1 1 <1F. odoratum 1 <1Pasteurella multocida 1 <1 1 <1Proteus inconstans 1 <1P. paucimobilis 2 <1A. hydrophila 135 53 114 41 138 40

a Each value indicates percent of total isolates.b Species not identifiable by available techniques.

available supplies of MacConkey and Endo me-dia.

DISCUSSIONThe comparative coliform tests were done on

samples of water collected from a wide varietyof sources. In these samples, the numbers ofcoliforms and bacteria which may interfere withcoliform counts varied from very high to 0, andthey included coliforms exposed to the riverenvironment, conventional wastewater purifica-tion and advanced tertiary processes such aslime treatment, sand filtration, active carbontreatment, and disinfection by means of ozoneand chlorine. Endo agar yielded the highest av-erage coliform counts for all these samples (Ta-bles 1 and 2). Teepol agar generally had highercounts than Teepol broth, whereas the lowestcounts were usually recorded on MacConkeyagar. Teepol broth was omitted from many com-

parative tests, because using the saturated padsproved inconvenient and time consuming, thepads tended to dry out, and the agar-based Tee-pol medium proved to generally yield highercounts. A similar experience has been reportedfor pads saturated with liquid Endo (38) andMacConkey (22) media. The slightly higher costof agar-based media is justified by more reliableresults, convenience and saving in time (22, 33,38).

Identification of 871 isolates from 13 samplesof various origins showed that A. hydrophilawas the species most frequently present amongcolonies which conformed to the definition oftotal coliform bacteria on MacConkey, Teepol,and Endo agars (Table 3). Endo agar provedslightly more specific than Teepol agar, andmuch more specific than MacConkey agar, forspecies of Escherichia, Citrobacter, Klebsiella,and Enterobacter, which is the group of primaryinterest in total coliform tests (37).

VOL. 38, 1979

on April 20, 2019 by guest

http://aem.asm

.org/D

ownloaded from

356 GRABOW AND DU PREEZ

Most probable number (MPN) tests were ex-cluded from this evaluation of total coliformmethods, since evidence has been presented thatMF is the technique of choice for general pur-poses (11, 20, 22, 34). The advantages ofMF are:it gives more accurate counts (11, 22, 34); it givesa direct count, whereas MPN evaluations arebased on statistical estimates with inherent er-rors (9, 35); it yields results within 16 to 24 h andMPN evaluations only after 48 to 96 h; it yieldsvaluable additional information on organismssuch as Aeromonas and Pseudomonas species(22) and even pathogens like Yersinia enteroco-litica (32, 42); colonies can easily be picked frommembranes for further identification; larger vol-umes of water are being tested in standard MFtests and the volume can easily be increasedextensively if necessary; organisms such as Clos-tridiumperfringens (9) and coliphages (43) mayinterfere with MPN evaluations; MF tests arecheaper than MPN evaluations (20, 22), andtheir performance is less cumbersome and timeconsuming; the petri dishes used for MF take upless incubator space than the racks with tubesrequired for MPN evaluations; and MF mayconveniently be applied under field conditions(48). MPN evaluations yield higher counts forchlorinated effluents than standard MF tech-niques without enrichment procedures (35, 39).If for any particular reason the higher counts onchlorinated effluents are required, the additionof relatively simple enrichment procedures toMF techniques will yield counts equivalent tothose of MPN evaluations (9, 35). The signifi-cance of these higher counts, which are attrib-uted to the inclusion of stressed coliforms (35),is uncertain. In chlorinated effluents, pathogensare presumably stressed to the same extent ascoliforms. There is no evidence that thesestressed pathogens are of any health signifi-cance, since they may not be able to survive hostdefense mechanisms. Standard MF tests for to-tal coliforms have rarely if ever failed to provethe microbiological (including virological) safetyof properly treated drinking water (1, 8, 17, 22,23, 24). The additional labor and cost of enrich-ment procedures are therefore not regarded nec-essary for general purposes and routine analysisof drinking water supplies (2, 22, 33). MPNevaluations remain useful for tests on highlyturbid samples which clog membranes. How-ever, the turbidity of drinking waters should bewell below levels which may affect the efficiencyof MF techniques (18).

Proposals to exclude cytochrome oxidase-pos-itive organisms, mainly Aeromonas and Pseu-domonas species, from total coliform counts (3,37) should be considered with caution. The total

APPL. ENVIRON. MICROBIOL.

coliform count should not be regarded as a spe-cific indicator of fecal pollution. The fecal coli-form count primarily serves this purpose (10).Even if an oxidase test is included, the totalcoliform count cannot meet this requirement,since many oxidase-negative coliforms multiplyin the water environment and are not of fecalorigin (4, 10). The total coliform count shouldbe regarded primarily as an indicator of thesanitary quality of drinking water (1, 17).Properly designed, operated, and controlleddrinking water plants consistently produce waterwhich is free of total coliforms per 100-ml samplewhen tested by standard techniques (1, 17, 18,21, 24). A positive total coliform test and thepresence of lactose-negative bacteria which ov-ergrow membranes and obscure coliforms, indi-cate inefficient treatment, secondary contami-nation, or aftergrowth, none of which should betolerated (1, 7, 21, 46). Coliforms isolated fromdrinking water should immediately be identifiedto establish possible fecal origin. In addition totheir indicator value, bacteria which yield a pos-itive coliform test may themselves constitutehealth hazards and should therefore also not betolerated in drinking water. K. pneumoniae,which readily multiply in various water environ-ments, constitute an opportunistic pathogen ofincreasing importance (4). A. hydrophila, whichprimarily multiply in water, are not only patho-genic to fish and various other animals, but mayalso infect humans (14, 31). Gram-negative bac-teria which multiply in highly purified watersuch as distilled water cause serious problems inhospitals (13). Bacteria involved in standard col-iform tests may also carry hazardous plasmidswhich are transferable among gram-negative or-ganisms. These plasmids include resistance fac-tors which confer on their hosts resistance toantimicrobial drugs, disinfectants, and variousother agents (27, 28), whereas others may codefor enterotoxin production which turns normallyharmless bacteria such as E. coli into a seriouspathogen (12, 29, 41). In view of these consider-ations, the oxidase requirement would unneces-sarily increase the cost, labor, time, and techni-cal know-how needed for the coliform test andat the same time limit its sensitivity, efficiency,and reliability.

Evaluation of the above advantages and dis-advantages of available techniques for total col-iforms indicates that the method of choice forroutine analysis of drinking water is MF, usinggood quality membranes such as those used inthis study (36, 45) and incubation at 35 + 0.50Cfor 22 to 24 h on m-Endo LES agar withoutenrichment procedures or an oxidase test. Thistechnique is not only reliable but also the most

on April 20, 2019 by guest

http://aem.asm

.org/D

ownloaded from

VOL. 38, 1979

appropriate for laboratories in small communi-ties or developing countries with limited funds,laboratory facilities, and trained staff.

ACKNOWLEDGMENTSThanks are due to Irmela G. Middendorff and J. S. Burger

for skillful technical assistance, to N. P. Nicolle, Chief Chemistof the Pretoria Municipality, for permission to sample theDaspoort purification works and the Apies River, and to 0.W. Prozesky and L. S. Smith for their advice.

This paper is published with the approval of the Directorof the National Institute for Water Research.

LITERATURE CITED1. Allen, M. J., and E. E. Geldreich, Jr. 1978. Evaluating

the microbial quality of potable waters, p. 3-11. In C.W. Hendricks (ed.), Evaluation of the microbiologystandards for drinking water. U.S. Environmental Pro-tection Agency, Washington, D.C.

2. American Public Health Association. 1975. Standardmethods for the examination of water and wastewater,14th ed. American Public Health Association, Inc.,Washington, D.C.

3. Aurand, K., U. Hasselbarth, G. Miller, W. Schu-macher, and W. Steuer. 1976. Die Trinkwasser-Ver-ordnung. Erich Schmidt Verlag, Berlin.

4. Bagley, S. J., and R. J. Seidler. 1977. Significance offecal coliform-positive KkbsieUa. Appl. Environ. Mi-crobiol. 33:1141-1148.

5. Bordner, R H. 1977. The membrane filter dilemma, p.8-11. In R. H. Bordner, C. F. Frith, and J. A. Winter(ed.), Proceedings of the Symposium on the Recoveryof Indicator Organisms Employing Membrane Filters.U.S. Environmental Protection Agency, Cincinnati,Ohio.

6. Bordner, R. H. 1978. Standardization of microbiologicalmethods for analysis of water. J. Food Protect. 41:314-319.

7. Burman, N. P., and J. S. Colbourne. 1977. Techniquesfor the assessment of growth of micro-organisms onplumbing materials used in contact with potable watersupplies. J. Appl. Bacteriol. 43:137-144.

8. Craun, G. F. 1978. Impact of the coliform standard onthe transmission of disease, p. 21-35. In C. W. Hendricks(ed.), Evaluation of the microbiology standards fordrinking water. U.S. Environmental Protection Agency,Washington, D.C.

9. Department ofHealth and Social Security. 1970. Thebacteriological examination of water supplies. Reportson public health and medical subjects no. 71. Her Maj-esty's Stationery Office, London.

10. Dufour, A. P. 1977. Escherichia coli: the fecal coliform,p. 48-58. In A. W. Hoadley and B. J. Dutka (ed.),Bacterial indicators/health hazards associated with wa-ter. American Society for Testing and Materials, Phil-adelphia.

11. Dutka, B. J., and S. E. Tobin. 1976. Study on theefficiency of four procedures for enumerating coliformsin water. Can. J. Microbiol. 22:630-635.

12. Echevema, P., L. Verhaert, C. V. Ulyangco, S. Ko-malarini, M. T. HO, F. Orskov, and I. Orskov. 1978.Antimicrobial resistance and enterotoxin productionamong isolates of Escherichia coli in the far east. Lancetii:589-592.

13. Favero, M. S., N. J. Petersen, L. A. Carson, W. W.Bond, and S. H. Hindman. 1975. Gram-negative waterbacteria in hemodialysis systems. Health Lab. Sci. 12:322-334.

14. Fliermans, C. B., R. W. Gorden, T. C. Hazen, and G.W. Esch. 1977. Aeromonas distribution and survival ina thermally altered lake. Appl. Environ. Micribiol. 33:114-122.

TECHNIQUES FOR TOTAL COLIFORMS 357

15. Frankel, S., S. Reitman, and A. C. Sonnenwirth (ed.).1970. Gradwohl's clinical laboratory methods and di-agnosis, vol. 2, p. 1117. C. V. Mosby Co., St. Louis.

16. Geldreich, E. E. 1977. Performance variability of mem-brane filter procedures, p. 12-19. In R. H. Bordner, C.F. Frith, and J. A. Winter (ed.), Proceedings of theSymposium on the Recovery of Indicator OrganismsEmploying Membrane Filters. U.S. Environmental Pro-tection Agency, Cincinnati, Ohio.

17. Geldreich, E. E. 1978. Bacterial populations and indicatorconcepts in feces, sewage, stormwater and solid wastes,p. 51-97. In G. Berg (ed.), Indicators of viruses in waterand food. Ann Arbor Science, Ann Arbor, Mich.

18. Geldreich, E. E., M. J. Allen, and R. H. Taylor. 1978.Interferences to coliform detection in potable watersupplies, p. 13-20. In C. W. Hendricks (ed.), Evaluationof the microbiology standards for drinking water. U.S.Environmental Protection Agency, Washington, D.C.

19. Geldreich, E. E., and N. A. Clarke. 1972. The coliformtest: a criterion for the viral safety of water, p. 103-109.In Proceedings of the 13th Annual Sanitary EngineeringConference, Urbana, Ill. U.S. Environmental ProtectionAgency, Cincinnati, Ohio.

20. Geldreich, E. E., and H. Kennedy. 1978. The cost ofmicrobiological monitoring, p. 1-20. In Proceedings ofthe National Science Foundation Seminar, Philadel-phia, Pa. National Science Foundation, Washington,D.C.

21. Grabow, W. 0. K. 1977. South African experience onindicator bacteria, Pseudomonas aeruginosa, and R'coliforms in water quality control, p. 168-181. In A. W.Hoadley and B. J. Dutka (ed.), Bacterial indicators/health hazards associated with water. American Societyfor Testing and Materials, Philadelphia.

22. Grabow, W. 0. K. 1979. Membrane filtration, SouthAfrican view. In B. J. Dutka (ed.), Membrane filtration:techniques, applications and problems. Marcel Dekker,New York, in press.

23. Grabow, W. 0. K., B. W. Bateman, and J. S. Burger.1978. Microbiological quality indicators for routinemonitoring of wastewater reclamation systems. Prog.Water Technol. 10:317-327.

24. Grabow, W. 0. K., and M. Isaacson. 1978. Microbiolog-ical quality and epidemiological aspects of reclaimedwater. Prog. Water Technol. 10:329-335.

25. Grabow, W. 0. K., L. G. Middendorff, and N. C. Bas-son. 1978. Role of lime treatment in the removal ofbacteria, enteric viruses, and coliphages in a wastewaterreclamation plant. Appl. Environ. Microbiol. 35:663-669.

26. Grabow, W. 0. K., and E. M. Nupen. 1972. The load ofinfectious microorganisms in the waste water of twoSouth African hospitals. Water Res. 6:1557-1563.

27. Grabow, W. 0. K., 0. W. Prozesky, and J. S. Burger.1975. Behaviour in a river and dam of coliform bacteriawith transferable or nontransferable drug resistance.Water Res. 9:777-782.

28. Grabow, W. 0. K., 0. W. Prozesky, and L. S. Smith.1974. Drug resistant coliforms call for review of waterquality standards. Water Res. 8:1-9.

29. Grabow, W. 0. K., M. Van Zyl, and 0. W. Prozesky.1976. Behaviour in conventional sewage purificationprocesses of coliform bacteria with transferable or non-transferable drug-resistance. Water Res. 10:717-723.

30. Hayek, L. J., and G. W. Willis. 1976. A comparison oftwo commercial methods for the identification of theenterobacteriaceae-API 20E and the enterotube-withconventional methods. J. Clin. Pathol. 29:158-161.

31. Hazen, T. C., C. B. Fliermans, R. P. Hirsch, and G.W. Esch. 1978. Prevalence and distribution of Aero-monas hydrophila in the United States. Appl. Environ.Microbiol. 36:731-738.

32. Highsmith, A. K., J. C. Feeley, P. Skaliy, J. G. Wells,

on April 20, 2019 by guest

http://aem.asm

.org/D

ownloaded from

358 GRABOW AND DU PREEZ

and B. T. Wood. 1977. Isolation of Yersinia enteroco-litica from well water and growth in distilled water.Appl. Environ. Microbiol. 34:745-750.

33. Inland Waters Directorate. 1978. Methods for micro-biological analysis of waters, wastewaters and sedi-ments. Canada Centre for Inland Waters, Burlington,Ontario.

34. Janardan, K. G., M. C. Long, and D. J. Schaeffer.1977. Recovery of coliforms by the MPN and MF tech-niques using a 2w-factorial experimental design. 11EQDocument 77/07, Illinois Institute for EnvironmentalQuality, Chicago.

35. Lin, S. 1973. Evaluation of coliform tests for chlorinatedsecondary effluents. J. Water Pollut. Control Fed. 45:498-506.

36. Lin, S. D. 1977. Comparison of membranes for fecalcoliform recovery in chlorinated effluents. J. WaterPollut. Control Fed. 49:2255-2264.

37. Lupo, L., E. Strickland, A. Dufour, and V. Cabelli.1977. The effect of oxidase positive bacteria on totalcoliform density estimates. Health Lab. Sci. 14:117-121.

38. McCarthy, J. A., J. E. Delaney, and R. J. Grasso.1961. Measuring coliforms in water. Water SewageWorks 108:238-243.

39. Mowat, A. 1976. Most probable number versus membranefilter on chlorinated effluents. J. Water Pollut. ControlFed. 48:724-728.

40. Robertson, E. A., and J. D. MacLowry. 1974. Mathe-matical analysis of the API enteric 20 profile register

APPL. ENVIRON. MICROBIOL.

using a computer diagnostic model. Appl. Microbiol.28:691-695.

41. Rosenberg, M. L., J. P. Koplan, I. K. Wachsmuth, J.G. Wells, E. J. Gangarosa, R. L. Guerrant, and D.A. Sack. 1977. Epidemic diarrhea at Crater Lake fromenterotoxigenic Escherichia coli. A large waterborneoutbreak. Ann. Intern. Med. 86:714-718.

42. Schiemann, D. A. 1978. Isolation of Yersinia enteroco-litica from surface and well waters in Ontario. Can. J.Microbiol. 24:1048-1052.

43. Schiemann, D. A., J. L. Manley, and M. 0. Arnold.1976. Coliphage interference in the recovery of coliformbacteria. J. Water Pollut. Control Fed. 48:533-539.

44. South African Bureau of Standards. 1971. Specifica-tion for water for domestic supplies. South AfricanBureau of Standards, Pretoria.

45. Tobin, R. S., and B. J. Dutka. 1977. Comparison of thesurface structure, metal binding, and fecal coliformrecoveries of nine membrane filters. Appl. Environ.Microbiol. 34:69-79.

46. Van Der Kooij, D. 1977. The occurrence ofPseudomonasspp. in surface water and in tap water as determined oncitrate media. Antonie van Leeuwenhoek J. Microbiol.Serol. 43:187-197.

47. World Health Organization. 1971. International stand-ards for drinking water. World Health Organization,Geneva.

48. World Health Organization. 1976. Surveillance ofdrinking-water quality. W.H.O. Monogr. Ser. no. 63.World Health Organization, Geneva.

on April 20, 2019 by guest

http://aem.asm

.org/D

ownloaded from