fi-l “.A _ II I “ I‘L‘ III III I I I I II .I III .I‘II II" I I I I III \ .‘I I! \ 'J I II II I I I \ II I I II ,I ‘I I ‘I I I I: ’2 ::i : it; I I I'I I \ I III I STUDIES WETH MED-IA RECENTLY AVMLABLE FOR U55 WITH THE MEMBRAJ‘E F ELTER IN THE ENUMERATEQ‘N C35 COLIFORM ORGANISMS IN WATER T339555 {or We Degree of M. 5. MESHIGAN STATE UNE‘JEaSETY Jchn .11 Panes 1957 STUDIES WITH MEDIA RECENTLY AVAILABLE FOR USE WITH THE MEMBRANE FILTER IN THE ENUMERATION OF COLIFOBM ORGANISMS IN WATER by JOHN J . PANES A THESIS Submitted to the College of Science and Arts of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE The Department of Microbiology and Public Health 195? AN ABSTRACT JOHN J. PANES A study was undertaken of two media recently proposed for use, in conjunction with the membrane filter technique, for the estimation of the coliform bacteria in water supplies. .The first, developed by McCarthy, was found to be inadequate for use with natural, untreated waters. Experi- mental work was done in an attempt to render it more suitable, but no improvement could be made, and it was concluded that the basic formula was not suitable for further development. The second, recently available commercially, and known as M-Endo broth M.F., gave coliform determinations at 37°C approximating very closely the estimates obtained by the con- ventional 5-tube decimal dilution procedure. Not all of the organisms which grow upon this medium are coliform, and it is considered that the ecology, distribution, and sanitary significance of the non-coliform flora must be established before the medium can be accepted for routine use. A direct method, using this medium with the membrane filter technique, for the estimation of Escherichia coli in water supplies was also attempted. The results obtained were not satisfactory, in that in a number of instances, samples known to contain E.coli gave negative results, and organisms other than E.coli were included in the counts obtained. 1 2 The opportunity was taken to compare lauryl tryptose broth with MacConkey's medium in the standard S-tube decimal dilution procedure for the estimation of the coliform density of water supplies. Both media gave essentially similar re- sults, although the addition of a pH indicator to the lauryl tryptose broth is recommended. ACKNOWLEDGMENTS The author wishes to express his gratitude to the W. K. Kellogg Foundation and to the Ministry of Agriculture, Fisheries and Food for making his visit to the United States of America possible. He also wishes to express his appreciation to Dr. w. L. hallmann for his guidance and counseling during the course of the work, and his thanks to Dr. F. B. Peabody and Mr. I. L. Dahljelm for their assistance in the proofreading of the manuscript. The author is also indebted to the personnel of the stockroom in the Department of Microbiology and Public Health who never hesitated in giving their coOperation. 11 INTRODUCTION PART TABLE OF CONTENTS I. THE USE OF EOSIN AND METHYLENE BLUE . . . . . . Review of the Literature . . . . . . . . . . II. THE USE OF pH INDICATORS . . . . . . . . . . . Review of the Literature . . . . . . . . ... Experimental and Discussion . . . . . . . . . III. THE USE OF BASIC FUCHSIN AND SODIUM SULPHITE. . Review of the Literature . Materials and Methods . . . water‘sambles . . . . . . Membrane filter technique MiP.Ni_determinations . . Cultural studies . . . . O 0 9 O o o O 0 O O 0 o O O O 0 O O O O O O O O O O O O O O O O O O O O O O O O O I O O O O O O O O 0 O 0 0 Results, Discussion, and Conclusions . . . . A. B. C. The compgbison of lauryl tryptose and MacConkeykbroths . . . . . . . . . . . I22? comparison of coliform determina- tions by the membrane filter technique and the standard diiutioniprocedure . . The comparison of the counts of E.coli obtained at UHUC by the membrane filter method with M-Endo broth MF and by using thebj-tube dilution procedure with confirmatigg_of tubesbpgsitive in the_p§esumptive tegt in 8.6.8. broth a tgflhoc . . . . . . . . . . . . The Present Status of the Membrane Filter in the Bacteriological Examination of Water . . REFERENCES . 111 Page 47 53 57 TABLE II. III. LIST OF TABLES PAGE Results Obtained Using the Decimal S-tube Dilution Technique with MacConkey's Medium and Lauryl Tryptose Broth . . . . . . . . . . . 37 The Comparison of the Number of Sheen- Producing Colonies Developing on M-Endo Broth M.F. Used with the M.F. Technique and the M.P.N. of Coli-Aerogenes Organisms Calculated from the Standard 5-tube Dilution Method . . . . . . . . . . . . . . . . #6 The Comparison of the Number of Sheen- Producing Colonies Developing on M— Endo Broth M. F. Incubated at #40 C Using the M. F. Technique and the Number of E. coli Calculated from the 5- tube Dilution Method with Confirmation at 4&0 C . . . . . . . 52 iv INTRODUCTION During the past five years membrane filter techniques for the quantitative estimation of the coliform group of organ, isms in water have been developed. A satisfactory filtering procedure has by now been well established and suitable appa- ratus and filter material are commercially available. The difficulty has been, and still is, in the formu- lation of a suitable medium upon which to place the inocula- ted membranes for the growth and enumeration of the coliform group of organisms. It is a simple matter to grow organisms of this group, but due to the limited surface area available it becomes necessary to suppress the growth of the rest of the water flora to prevent the coliform organisms, especially if present in small numbers, from being overgrown and inhib- ited. It is also necessary to color the colonies in some way to render them visible against the white surface of the fil- ter. Black-colored filtering material is now available, but uDon autoclaving it becomes excessively brittle and difficult to handle. Exposure to ethylene oxide vapor or to ultra- violet light may be used satisfactorily, but these operations are somewhat cumbersome and are not suitable for routine use. The dyes and stains used have been such that the colonies are cxflored and those of the coliform organisms differentiated on 1 2 a color basis from those of other organisms capable of grow- ing upon the medium used. Gas production from lactose is used in the various tube-dilution methods as one of the criteria upon which the recognition of the coliform group of organisms is based, and indeed, by definition, the ability to produce gas from lac- tose is one of the most important diagnostic characteristics of the group. Since there is no possibility of being able to assess gas production from coliform organisms growing up- on the filter surface, other distinguishing properties must be utilized in their differentiation. The recognition of coliform organisms when grown upon solid media has been carried out for many years, generally by one of three methods: 1. The production of colonies of characteristic appearance upon a medium containing eosin and methylene blue. 2. By the use of a pH indicator in the medium which contains lactose, enabling those organisms which produce acid from this sugar to be recognized. 3. The production of colonies of characteristic appearance upon a medium containing basic fuchsin and sodium sulphite. All three of these systems of differentiation have been utilized in the media developed for use with the mem- brane filter and will be discussed further. PART I THE USE OF EOSIN AND METHYLENE BLUE Review of the Literature Eosin-methylene-blue agar was developed by Levine in 1918 (19) from a formulation by Holt-Harris and Teague (12) who used it as a plating medium following enrichment for the isolation of enteric pathogens. The medium was incorporated into the 1923 edition of Standard Methods (38) and remains the solid medium most widely used for the isolation of coli- form organisms. Descriptions of coliform colonies as they appear on this medium are given by Levine (20) and by Howard and Thompson (13). The composition of the medium is as follows:' Bacto-peptone . . . . . . . . 1.0 g. . Lactose . . . . . . . . . . . 1.0 g. Dipotassium phosphate . . . . 0.2 g. Agar . . . . . . . . . . . . 1.5 g. Eosin I . . . . . . . . . . . 0.0“ g. Methylene blue . . . . . . . 0.0065 g. Adjust pH to 7.1 The mode of action of E.M.B. agar is considered by Wynne et al. (51) to depend upon two factors: the formation by eosin and methylene blue of a methylene blue eosinate, *Ingredients per 100 m1. 3 I; made up of one molecule of each of the dyes, which acts as an acidic dye, and the production by the bacteria of a low pH which enables the cells of the colony to take up the acidic dye. Colonies of bacteria which do not ferment lac- tose are colorless because their cells, in the alkaline en— vironment maintained by them, do not take up the stain. levine (21) has shown that a sugar-free base is necessary in the preparation of the medium if differentiation of coliform types is required because the addition of fermentable carbo- hydrate causes all coliform colonies to take on a metallic sheen. Prescott giiai, (32) have reviewed the literature concerning the specificity of this medium. Identification of colonies is subject to an error, usually less than 10%, and which is on the safe side, false positive reactions be-- ing encountered with a much greater frequency than false negatives. Standard Methods (39) in accepting confirmatory identification purely by the appearance of colonies typical of Escherichia coli recognizes the accuracy of the medium in this respect. 0n the other hand, in requiring the complete identification of two or more colonies most likely to be coliform if typical E.coli are not present, the inability of the medium to indicate with satisfactory accuracy the pres- ence of coliform organisms other than E,coli is recognized. It should be noted that apart from the dyes used no other inhibitory agents are present, but even so, Mallmann 5 and Derby (23) have suggested that the medium may not support the growth of all coliform strains. Attempts have been made to utilize this type of medium in conjunction with the membrane filter procedure for the enumeration of coliform Organisms in water. In 1952, the U.S.P.H.S. Environmental Health Center, (E.H.C.) (47) pub- [A‘fl lished details of the preparation of a medium having the I following composition:* i Peptone (Albimi M) . . . . . . . . . . “.0 g. 1 Yeast autolysate . . . . . . . . . . . 0.6 g. Dipotassium hydrogen phosphate . . . . 0.7 g. SOdium Chloride O O O O O O O O O O O o .5 g. F ‘ Bacto-bile salts no. 3 . . . . . . . . 0.1 g. Lactose . . . . . . . . . . . . . . . 2.0 g. Methylene blue . . . . . . . . . . . . 0.1 g. Eosin Y . . . . . . . . . . . . . . . 0.5 g. Bile Salts are included in order to inhibit non- coliform organisms, and a 2+hour preliminary enrichment of the cells obtained by pre-incubating the membranes upon pads soaked in lactose broth is recommended before transfer to the differential medium. In addition, the filter material used must be dyed with methylene blue before filtration of the sample takes place. Even so, coliform colonies do not show their characteristic appearance, apparently due to the fact that the methylene-blue-eosinate complex cannot diffuse through the pores in the filter material (17). A similar formulation, with the exception that the concentration of bile salts is doubled, is given in the Technical Bulletin (#1). *Ingredients per 100 m1. 6 Neither has received much attention due, no doubt, to the atypical appearance of the colonies and the extra manipula- tion involved in the dying of the filter material. .8‘ mm" fin PART II THE USE 0? pH INDICATORS Review of the Literature These have been incorporated into liquid broths and solixi plating media used for the recognition of coliform Orgauaisms ever since the beginnings of water bacteriology. The broth developed by MacConkey (22) is still used in ,BPiIaain for the enumeration of coliform organisms in water, and. the same medium solidified with agar has been of value in the isolation and differentiation of coliforms and enteric Pathogens. Lactose is provided as a source of fermentable canohydrate and the presence of organisms fermenting this sugar is recognized‘by the change of color in the medium due to the production of acid. In the case of MacConkey broth, inVerted Durham fermentation vials are used to detect gas pI‘OCIuction. Until fairly recently, the indicator employed "as neutral red, but this has now been largely replaced by 'btfiom-cresol-purple which gives a sharper color change and 1"£119. been shown to be less toxic, especially to the strepto- °°cci (1). Taylor (#2), and Taylor gi_§i, (#3) have utilized a triple-strength MacConkey broth containing twelve times the 7 8 normal amount of brom-cresol-purple in conjunction with the membrane filter technique for the enumeration of coliform organisms in water. Later these authors (an) reported further progress in the study. Using the membrane filter procedure tliey compared this medium, which requires no enrichment, and tlie EHO modified Ehdo medium described later with a S-tube dilution technique in which primary incubation was at “2°C. 111 both cases the two membrane filter media gave higher counts but, as the authors point out, the high incubation temperature of the dilution procedure used may well have Prevented the growth of some coliform organisms. Goetz and Tsuneishi (9) have described the use of a Incxiified Gassner (8) medium for the same purpose. The for- lnila contains lactose as a source of fermentable carbohydrate 81nd, with the exception of the peptone present, contains no thiffer. TWo dyes are used: a pH indicator, ink blue, which is blue in acid; and a yellow dye, fast mordant yellow, which partially inhibits cocci and spore-formers, but does not af- fect the development of salmonellae and coliform organisms. IPre-treatment of the filter material with the yellow dye is necessary. The acidity produced by the fermentation of the lactose stains the colonies and their immediate environment ‘blue, while other colonies appear greenish-yellow in color. The addition of an oxidation-reduction indicator is recom- mended to distinguish the slow from the rapid lactose fer- menters, since the latter will turn blue before the redox 9 potential has dropped low enough to give the colonies a red ccxlor. 0n the other hand, the slow lactose fermenters will tirrn.red and will not develop the blue color appreciably. No results obtained with this medium are given, and it does :nxrt appear to have been used subsequently by other workers. (McCarthy (26), realizing the toxicity of the modified type of Endo medium described later, formulated a MacConkey type of medium for use with the membrane filter procedure in the enumeration of coliform organisms in water. The selec- tisve action depends upon the use of bile salts and a low ccmncentration of ethyl violet, and the recognition of lac- twose fermenters is facilitated by the inclusion of a pH indi- cmltor brom-cresol-purple. Methylene blue is included to color the filter material in order to provide a background against which colonies can be seen more easily. After 1h Ynours incubation at 35°C in an atmosphere saturated with vuater vapor, coliform colonies are colored yellow or pinkish- yellow and those of non-coliform organisms blue. The formula is given belowz" Bacto-tryptose . . . . . . 1.5 g. SOdium Chloride o o o o o 005 go Lactose . . . . . . . . . 1.0 g. Bacto-bile sal s . . . . . 0.5 g. Brom-cresol-purple . . . . 0.15 g. Methylene blue . . . . . . 0.001 g. Ethyl violet . . . .‘. . . 0.00175 3. With both samples of well and river water counts were obtained which were higher than those on the EEC modified ”Ingredients per 100 ml. medium 10 Endo medium and which approximated those calculated from the standard 5-tube dilution procedure with subsequent confirma- tion of positive tubes in 2% brilliant green bile broth. In the case of the river and highly polluted well waters, how- ever, there was often a spreading overgrowth of the filter surface due to the presence of non-coliform organisms, and the medium could not be regarded as satisfactory for the examination of natural raw water supplies. Experimental and Discussion It seemed probable that some improvement could be made in the medium, and a good deal of time was spent in an attempt to render this medium more satisfactory. Bacto-bile salts no. 3 were employed in lieu of the ordinary prepara- tion and a range of concentrations tested to determine the Opt 1mal amount required. In addition, the effect of bril- liant green and sodium lauryl sulphate was determined over a Wide range of concentration for each material both singly and in combination with each other and the bile salts. This was accomplished by inoculating a number of membranes with a p0lluted water known to contain both E.coli and Aerobacter \aePOgenes according to the filtering procedure presented in St"Elndard Methods (39). The equipment and filter material we're obtained from the Millipore Filter Corporation.” A \— *Watertown, Mass . ll basal medium containing the tryptose, sodium chloride, brom- cresol-purple, and methylene blue was prepared in concentra— ted form. Solutions of suitable strength of lactose, ethyl violet, brilliant green, and the bile salts no. 3 were pre- pared and all were sterilized separately in the autoclave at 121°C for 15 minutes. The complete medium was prepared from the components in the concentrations under test and the volume adjusted with sterile distilled water to a constant level, the con- centration of base remaining constant. Sterile absorbent Pads in sterile 50 mm. Pyrex petri dishes were wetted with a fixed amount (1.7 ml.) of the medium, the inoculated mem- branes rolled upon their surfaces and incubated in the in- verted position in an atmosphere saturated with water vapor at 35°C for 1h - 18 hours. It was found that a series of membranes could be inoculated successively from one water sE-‘tunple without the need for resterilization of the apparatus ea-ch time, provided that the entire inner surface of the glass container in contact with the sample. was rinsed with 30 m1. of sterile water between each inoculation. 0n thirty Occasions a second sterile membrane was inserted and the QCDritainer again rinsed thoroughly with sterile water. When 1heubated upon the coliform medium no colonies developed, showing that there is no danger of carrying over coliform oT'ganisms from one filtration to the next. 12 No improvement over McCarthy's original formula could be made -- the brilliant green, ethyl violet, and sodium lauryl sulphate bringing about little improvement in the restriction of colony size and inhibition of non-coliform or‘Sanisms, except in concentration sufficient to reduce the viable coliform population considerably. It appears that the failure of this medium lies in its inability to inhibit the growth of non-coliform organisms and to restrict the growth of the coliforms to the extent that their colonies remain discrete, while at the same time allow- ing all coliform cells deposited upon the filter surface to develop into colonies. Ethyl violet, brilliant green, and S“Odium lauryl sulphate are known to be most active against gram-positive organisms, and their ineffectiveness indicates t”lat the interfering organisms are gram-negative. A large 1ill-Amber of non-coliform colonies were picked from the filter 8ut‘faces and were shown to be decolorized in the gram stain- 111% procedure. It also became evident that a concentration or bile salts which would inhibit these organisms would also pr‘event many coliform cells from growing upon the filter sur- fafie; and in all probability this effect would be enhanced by their suspension, and consequent attenuation, in water for long periods. Such inhibitory effects have been overcome to some eRtent by the preliminary incubation of inoculated filters for a period of 2 hours upon a medium such as lactose broth 13 containing no inhibitory agents. _ A modification of this method which is described later involves the use of two ab- sorbent pads, superimposed upon each other, the upper being 1mPregnated with a plain lactose broth and the lower with the selective medium. During the time taken for the selec- tive medium to reach the cells they have been resuscitated by exposure to the lactose broth and are better able to with- stand the toxic effect of the inhibitors used in the selec- tive medium. A modification of this method was tested in Which the medium described was prepared in the form of a 1.5% agar gel, cooled to hsoc, and a layer superimposed upon another solidified agar medium containing the nutrient base and larger amounts of brilliant green, sodium lauryl sulphate, and bile salts. It was considered that the diffusion of the 1111'11bitors in the lower medium through the upper layer to the filter surface would take place slowly enough to permit 1211 tial colony development but would later restrict the growth of the coli-aerogenes bacteria and inhibit the growth Each inhibitor in the lower Dif- °f other organisms present. med ium was tested over a wide range of concentration. rePent thicknesses of surface medium, restricted only by the phlrsical limitations of the 50 mm. petri dish used, were also 811‘played. It appears that the inhibitors used, if present in Sufficient amounts, are able to diffuse through the agar rapidly enough to completely inhibit colony development, the ei‘f‘ect being identical to that of an homogenous medium 1“ containing larger amounts of inhibitors. It has been ob- served that, especially in the case of the Endo type media referred to later, colony develOpment upon the filter sur- face is not a gradual process but rather that there is a long lag period extending for about 10 hours during which no growth is visible. Subsequently, growth is extremely vigor- ous, and during the next few hours colonies of l - 2 mm. in diameter develOp. Presumably the 10-hour period is suffi- cient for the inhibitors in the lower medium to diffuse to the filter surface and to prevent subsequent colony develop- ment. It appeared that further study of this type of medium would not be profitable, and attention was directed to the Endo type of medium described in Part III. PART III THE USE OF BASIC FUCHSIN AND SODIUM SULPHITE Review of the Literature Endo, in 1904, (5) proposed a medium for the isolation of typhoid organisms consisting of nutrient agar containing lactose and an indicator system composed of basic fuchsin and sodium sulphite. It is still widely used as a plating medium for coliform organisms and is recommended for this purpose in Standard Methods (39) which gives details of its composition and preparation. The formula is reproduced below:* Dipotassium phosphate Bacto-peptone . . . . Agar . . . . . . Lactose . . . . . Sodium sulphite . Basic fuchsin . . The virtue of this medium tem. 'E.coli colonies develOping 0035 80 1.0 g. 2.0 g. 1.0 g. 0.25 g. 0.05 g. lies in its indicator sys- on it produce dark red, flatm‘button-like colonies with a distinct metallic sheen, "“1163 other eoliform types are nucleated, pink or red in color-, and possess no such luster. A great deal of atten- t1Oil-has been paid to the amounts of indicator and sulphite # “Ingredients per 100 ml. medium 15 16 required, and they have been found to be critical. Care is also needed in selecting the source of the fuchsin as not all lots are suitable. Endo attributed the appearance of the colonies on this medium to acid formation, but this alone does not pro- duce the deep red color and sheen typical of E.coli colonies. Other workers believed the reaction to be similar to that of Schiff's reagent with aldehyde, but the two are not appar- ently identical. DeBord (a) considered the effect to be caused by the combined effect of acid and aldehyde, since in his experiments neither alone could produce the characteris- tic color effect. Margolena and Hansen (25) concluded from their study that the effect was best eXplained by the theory <3f Neuberg and Nord (30) who postulated that the sulphite ‘traps small amounts of acetaldehyde, an intermediate product cof fermentation, as it is produced by coliform organisms; zind the medium then acts as a reagent to indicate the pres- ence of accumulated products. Thus it would appear that the ciiagnostic characteristic of gas production from lactose is replaced by the production of acetaldehyde. In the search for a suitable medium for the enumera- txion of coliform organisms in conjunction with the membrane filter technique, it was natural that attention should be given to this medium of Endo which has proved to be so use- fhxl as a differential medium in the past. l7 Schutz and Kruse (36) credit the Russians with using it in this connection in 1934, when they incubated inocula- ted membranes upon an Endo type of agar. The application of this type of medium to the mem- brane filter technique was initiated in America by the ILS.P.H.S. Environmental Health Center (E.H.C.) (2, 15) punch in 1951 published details of the preparation and use of EHC modified Endo medium, the composition of which is as follows:* Lactose . . . . . . . . 2.0 g. Dipotassium hydrogen phosphate . . . 0.7 g. Neapeptone . . . . . . . 2.0 g. Sodium sulphite (anhydrous) . . . . 0.27 g. Basic fuchsin . . . . . . . . . . . 0.09 g. Ethyl alcohol . . . . . . . . . . . 1.5 ml. Basic fuchsin itself is inhibitory and plays a part 111 regulating the size of colonies developing on the filters. TTue only other material present likely to inhibit growth is the ethyl alcohol used as a solvent for the fuchsin. It is Iptwobable that the concentration used is sufficient to bring about some reduction in spreading and to decrease colony size since Floyd and Dack (7) have shown that double this amount is effective in inhibiting the spreading of Proteus. Nevertheless the authors found that this medium was ixfliibitory to coliform organisms, a high prOportion of those present in fresh sewage failing to grow. In order to over- come this difficulty, a preliminary 2-hour enrichment, __ ”Ingredients per 100 ml. 18 obtained by placing the inoculated membranes upon pads soaked in lactose broth prior to incubation on the differ- ential medium, was necessary and resulted in a doubling of the sheen-producing colonies obtained. With this procedure a good agreement was found between coliform counts so ob- tained and those resulting from the use of 5-tube M.P.N. determinations with subsequent confirmation according to Standard Methods (39). It has been pointed out (15) that the classical de- scription of coliform colonies on Endo agar does not apply to those developing on the membrane filters using the modi- fied EHC procedure, since on the latter all of the E.coli and most of the A.aerogenes and intermediates develop the metallic sheen. Clark and Kabler (3) further state that a few of the A.aeroggn_e_s_ and intermediates may be missed by this procedure, but consider that the increased accuracy obtained compensates for this small loss. Neumann (31) examined several different types of media for the same purpose and found this EHC procedure the most satisfactory but not entirely suitable for all of the water sources tested, since other organisms of no known sanitary Significance were not inhibited. Kabler and Clark in 1952 (17) also considered this medium to be the best available at that time. Streicher (40) in 1951 suggested that as had been found with the original Endo medium the fuchsin : sulphite 19 ratio was of great importance. Later Clark and Kabler (3) and the workers at the Environmental Health Center (#8) con- firmed this view and published methods for titrating the in- dicator system to give more satisfactory and consistent results. Goetz gi__a_]_._. (10) have also attempted to correlate this membrane filter procedure using EHC Endo media with M.P.N. determinations using the 5-tube dilution technique described in Standard Methods (39). Only a very few sets of results are presented but fairly good agreement was obtained. A modification of the EHC technique is presented employing What the authors describe as a nutrient schedule procedure Which is less time-consuming and gives rather better results. In the method described the filter is incubated upon a pad made up of a thin leaf of porous paper containing a lactose broth above a heavy disc containing the differential Endo medium. The two pads, after being impregnated with their reapective media, are dried, placed together and sterilized, and require only moistening before use. The principle is trust the inhibitory materials in the lower pad will not reach the attenuated coliform cells until they have been resuscitated and are more resistant to their effects. The method, despite the advantages claimed by its originators, does not appear to have been much used by later workers. Presnell M. (33) using sea water also compared the two methods of enumeration using the 95% confidence 20 liJnits of the M.P.N. determinations developed by Velz (#9). Briefly this takes into account the bias of 5-tube M.P.N. determinations. By multiplying the M.P.N. obtained by fac- tors of 0.3 and 2.9, a range is obtained. It is considered truat.any coliform count obtained with the membrane filter iflYL10h falls within this range is not significantly different. By this manipulation 87% of the samples examined were in agreement. Thomas and Woodward (#5) have analyzed statistically tflae results obtained by three groups of investigators (:33, 52, 50) who have compared the same two methods of enu- meration, the M.P.N. determinations again being assessed on the basis of the 95% confidence limits. ‘Ihey concluded that the 5-tube dilution technique as presented in Standard Methods gives results higher than those obtained with the EHTC medium in conjunction with the membrane filter by a fac- tcrr of 1.0 - 1.9 with an average of 1.3. In addition, their iarualysis indicated that the two methods of enumeration do not measure the same groups of organisms. The differences were considered to be minor and in the opinion of the authors tkuere was no reason why the S-tube dilution method should be favored more than the membrane filter technique described. In 1955, Thomas g£_§;. (#6), using finished waters, carwfiled out further experiments and statistical analyses °°mpering the degree of control which could be exercised upon such waters using the two methods of enumerating 21 coliform organisms, and concluded that the membrane filter method could give equally good, if not better, control of treatment processes. The first qualitative report on the types of colonies found upon filter surfaces incubated on EHC modified medium was published by Jeter 9_t_§_l_. (14) who in addition typed, using the IMViC tests, a large number of the organisms con- cerned. They found that of 370 sheen-producing organisms isolated, only 92$ produced gas from lactose broth; but, even so, coliform counts obtained were lower than those obtained by the conventional 5-tube dilution methods with subsequent confirmation. This discrepancy, however, was not due to the exclusion from the Endo medium of one particular IMViC type as all were represented. In addition, of 361 colonies not Producing sheen, 4% produced gas in lactose. Again all IMViC types with only one exception (+ - + 4) were represented. Mention is also made of sheen-producing colonies which upon subculture would produce no gas from lactose broth in 48 hours at 35°C and which were, of course, counted as coliform. Thus it appears that the medium is not absolutely satisfac- tOPY in differentiating coliform organisms, and it seems that the ability to produce gas from lactose is not correlated Per‘fectly with the property of forming acetaldehyde. Nothing is known about the distribution and sanitary significance of these non-coliform, sheen-producing organisms; but, obviously, it Present in large numbers in a water supply, a false 22 assessment would be arrived at using this membrane filter method. This work tends to support that of Thomas and Woodward (#5) which has already been reviewed and who con- eluded that the two methods of enumeration are not measuring the incidence of the same groups of organisms. In order to overcome some of the difficulties and discrepancies discussed, various modifications of the tech- nique and media used have been suggested from time to time. Kabler (16) used the same formula as that described above except that the lactose in the modified Endo medium was autoclaved separately and the dye content reduced slight- 13'- During the period 1952 - 1953 he examined 1,706 samples of water using this medium with the membrane filter tech- nlque, comparing the results obtained with the standard S-tube dilution test, again using the 95% confidence limits of the M.P.N. Of his samples 7h% showed agreement, although as Pointed out by McCarthy (26), if completely negative samples, 1.3 . those showing no colifgrms by either method, are omit- ted . only 61% are in agreement. He also found that in eighty- f1Ve samples, coliform organisms could be detected by the 8t‘a‘l‘ldard tube-dilution method, none being observed upon the ‘ membrane filters. ‘ In addition, two samples showed the pres-D ence of coliform organisms by the membrane filter technique .- and gave zero M.P.N. determinations. It would seem that coli- f°Pm organisms are being missed by this modification of the' met"bod. 23 Levin and Laubausch (18) reported on the testing by both methods of 207 samples of drinking water delivered to "Hotel“ trains under emergency conditions. Of these three showed the presence of coliform organisms with the confirmed tube—dilution method, two of them producing coliform-like colonies upon the membrane filter. These authors also auto- claved the lactose in the Endo medium separately, and in an attempt to render the medium less inhibitory, reduced the dye—sulphite content to 0.7 of the amount producing optimal results in a titration procedure. One of the difficulties experienced in the counting of coliform colonies develOping upon the filter surface is that many non-coliform organisms also grow and if present in large enough numbers may crowd out the coliforms present. In an attempt to reduce this "background" population, Yee Eli—9‘1- (52) used oxine (8-hydroxyquinoline) as an inhibitor for‘ these organisms. Details of the composition and prepar- ation of this medium, known as LCC Endo, may be found in the Technical Supplement (’41) distributed by the Millipore Fllher Corporation. The results obtained were not consis- tent , a reduction in non-coliform population being found °n1y with certain .waters. In addition, the authors found that only 81% of sheen-producing colonies on this medium pr°d~uced gas from lactose and that 22% of non-sheen produc- lng types did ferment the sugar with the production of gas. A statement is made that if the filter surface is allowed to 2“ dry out after incubation, then more colonies will develop the characteristic sheen, but no evidence is presented that these are coliform organisms. The addition of brilliant green known to be particu- larly effective against gram-positive organisms‘has also been recommended for reducing the numbers of non-coliform organisms growing upon the filter surfaces and is suggested in Standard Methods (39) for this purpose in conjunction With the EHC modified Endo medium. No published data appear to be available concerning its performance. Hajna and Damon (11) used 0.02% sodium desoxycholate in order to bring about the same effect in a medium known as H-HD Endo broth (#1). It utilizes fuchsin and sulphite as the indicator. Yeast extract, casitone, and thiopeptone are also included, presumably to overcome the toxic effect of the fuchsin and desoxycholate against attenuated coliform or‘ganisms. No enrichment is required. Ethyl alcohol in a concentration of 1.5% is also present. The authors state that false positive reactions (the presence of sheen- pr'Oducing non-coliform colonies) are eliminated, but no °°mparison with M.P.N. determinations was made. The method "as used only qualitatively to detect the presence or ab- senee of coliform organisms. Out of 147 samples examined by this method and the standard 5-tube confirmed M.P.N. proce- dupe . fifty were positive by both methods and seventy were negative. In eight samples the membrane filter method gave 25 negative results, while coliforms were found by the tube- dilution test, and the reverse was true in seventeen cases. McCarthy (27) later carried out further studies using the same medium but on a quantitative basis. He found that it gave good results combined with the membrane filter tech- nique in. the examination of samples of diluted fresh sewage. Comparisons were made using the standard 5-tube dilution method. When water samples were used, a rather different picture was obtained, the counts with membrane filter method being considerably lower. In addition, he found that con- siderable numbers of organisms not producing the character- istic sheen upon this medium were, in fact, coliform; and by picking off all of the colonies on a particular filter into lactose broth,. he was able, -on the basis of gas production in that medium, to double his original coliform counts. If incubation was continued for a total period of 48 hours (at Presumably 35.500) a doubling of the count also resulted, and he concluded that the medium left much to be desired. In the same year Slanetz and Bartley (37) described the use of another modified Endo medium which does not re- quire preliminary enrichment and which is known as M-Endo “nth. the formula of which follows” Again yeast extract and either neopeptone or thiotone are °mployed, and no inhibitors apart from the fuchsin- sulphlte indicator are present. \ ”Ingredients per 100 ml. medium 26 Yeast extract . . . . . . . . . . . 0.6 g. Neopeptone . . . . . . . . . . . . . 2.0 g. Lactose ... . . . . . . . . . . . . 2.5 g. Dipotassium hydrog n phosphate . . . 0.7 g. Basic fuchsin . . . . . . . . . . . 0.1 g. Sodium sulphite . . . . . . . . . . 0.25 g. The medium was compared with others available for use with the membrane filter at that time, 1.8. EHC modified Endo broth (41) , the M-HD Endo broth described above, as well as the triple strength MacConkey broth suggested by Taylor _e_E_a_1_. (43, M4), and was found to be superior. Incu- bation at 35 or 37°C gave similar results, and a 22 - 21+ lunxr .incubation period was required. In contrast to the ex- perience of other workers, incubation in an atmosphere satu- ratexi with water vapor was not found to be necessary. Fifty-five samples from eight rivers, one brook, and one pond were examined, and counts of colonies showing sheen obtained on this [medium in conjunction with the membrane filter technique wePetcompared with the results of the standard 5-tube dilu- tion procedure followed by confirmation of positive tubes in 25‘ brilliant green bile broth. While the membrane filter °°uhts were generally lower than the M.P.N. determinations, g°°d agreement was obtained if the 95% confidence limits of the latter were used in the assessment. Cultural studies of 193 sheen—producing colonies were undertaken, and 97% proved to be coliform. The authors concluded that this medium allied with the membrane filter procedure was an efficient method for the detection and enumeration of coliform bacteria in water supplies . 27 In addition to the media listed above, these authors, in an effort to reduce the numbers of non—coliform organisms growing upon their filters, tested the performance of a num- ber of inhibitory agents in this respect. Penicillin, 8- hydroxyquinoline (oxine), and brilliant green were tested, but none were satisfactory. Since these materials are known to be particularly effective against gram-positive organisms, it would appear that the interfering flora growing upon these filters is predominantly gram-negative, and this is borne out by the experience of the author. Recently a further modification of Endo's medium has ‘beeri marketed by the Millipore Filter Corporation. This is knomna as M-Endo Broth MFR, and is manufactured by Difco Lab- orainories. Again, no preliminary enrichment is required. The composition is as follows:‘E Bacto-yeast extract . . . . . . 0.15 g. Bacto-casitone . . . . . . . . 0.5 g. Bacto-thiopeptone . . . . . . . 0.5 g. Bacto-tryptose . . . . . . . . 1.0 g. Lactose O O O O O O 0 O O O O O 1 .25 g. Sodium desoxycholate . . . . . 0.01 g. Dipotassium phosphate . . . . . 0.4375 g. HonOpotassium phosphate . . . . 0.1375 g. Sodium chloride . . . . . . . . 0.5 g. Sodium lauryl sulphate . . . . 0.005 g. Sodium sulphite . . . . . . . . 0.21 g. Bacto-basic fuchsin . . . . . . 0.105 g. It will be noticed that the fuchsin - sulphite content is alulost the same as in those media previously described but that? two inhibitors, sodium desoxycholate and sodium lauryl \ *Ingredients per 100 ml. .. '1‘1 I! ".n\| 28 sulphate, have been included, presumably to inhibit the growth of non-coliform organisms. Both of these materials are known to be effective against gram-positive organisms, but it has been demonstrated that the preponderence of non- coliform organisms growing upon the filter surface are gram— negative. No less than four complex sources of nitrogen are included, presumably to supply growth factors and micronutri- ents. It is difficult to conceive that all of these materi- als are necessary and no information is available regarding their inclusion. No results have been published concerning the use of this medium for the enumeration of coliform organisms in water supplies. Fifield 213;. (6) have utilized it to count col iform organisms, using the membrane filter procedure, in pasteurized milk, comparing counts so obtained with plate counts on desoxycholate agar. Good agreement was obtained between the two methods, but it is extremely doubtful whether, with all the extra manipulation involved, the membrane filter teckulique is suitable for this purpose. Preliminary work using pure cultures has been carried out (29). Coliform cultures of eight IMViC types were dilu- ted and suitable portions passed through the membrane filter discs and plated on nutrient agar. After incubation at 37°C for 21+ hours, colony counts were performed. Five sets of results are available for each culture. Very good agreement was obtained between these two methods, the counts obtained 4F" 29 with the membrane filter medium tending to be slightly lower, but no inhibition with the active cultures used was notice- able. This shows considerable improvement on the results obtained by McCarthy (26) with the EEC medium following pre- liminary enrichment and using similar methods. She used only one pure culture of E.coli, but found that a recovery of only 76% was obtained with this medium when compared with the col- ony count on brain-heart-infusion agar. On the basis of these results it was decided to find out how this medium would perform when used to assess the coliform content of water supplies. The advantages inherent in the membrane filter tech— nique are considerable and have been discussed and evaluated at some length by a number of workers (10, 36, 1&6). The ap- plication of the procedure for the estimation of the coliform Population of water supplies as indicated above, has not been successful, except perhaps in the experience of Slanetz and Bartley (37). This is reflected in the tentative approval of the technique and BBC modified media in Standard Methods (39) Which approves the method only under conditions where it has been found to give results comparable to the approved tube- dilution methods. The author is concerned with the bacteriological exaI'llination of small rural water supplies including wells, St"realms, Springs, boreholes, and rainwater, and it is with Such natural waters that present media have been found to be 30 inadequate. If a suitable medium were available, much time and materials could be saved by utilizing the membrane filter technique for this purpose, and it was with these facts in mind that the potentiality 'of the new medium was investiga- ted . Materials and Methods .s Water samples. As many different types of natural raw water supplies were examined as possible, those known by Previous testing to contain coli-aerogenes bacteria being selected for study. They comprised a number of driven and l drilled-wells, two rivers, and several lakes. Samples (were taken in 30-ounce bottles, fitted with screw caps and plas- tic liners, which had been sterilized in the autoclave at 121°C for 20 minutes before use. This provided a single 8aKiple sufficient for all tests to be performed. It was well Shaken before being tested -- always within 6 hours of sam- P1 ing. Membranegfglter techggggg. The glass filtration ap- Par‘atus, filter material, and absorbent pads were those sup- plied for the purpose by the Millipore Filter Corporation.fl F111“nation was carried out according to the procedure in the Tenth Edition of Standard Methods ( 39), the amount of sample flltered being adjusted to give less than 200 sheen-producing \ *Watertown, Mass. 31 colonies per membrane. Duplicate inoculated membranes pre- Pared from each sample were rolled upon the surface of ab- sorbent pads moistened with 1.7 ml. of the medium described and contained in 50 mm. Pyrex petri dishes. The medium used was Obtained in dehydrated form and was made up according to the instructions of the manufacturer.“ A sufficient amount was Prepared to meet the requirements for one day and was , used within 6 hours. The requisite amount was weighed out on Paper and transferred to a sterile 1-ounce screw—capped b(Hitle. The necessary quantity of 1.5% ethyl alcohol in sterile distilled water was added and the powder dissolved {— at “5°C before being heated in the steamer for 10 minutes. witShout this precaution, a black insoluble material was found to be present in the final product. Petri dishes containing the filters resting upon their absorbent pads were inverted and placed upon a perforated tray within a plastic box, the bottom of which was lined with wet cheesecloth. By this means an atmosphere saturated with water vapor was obtained. The box was placed in a “walk-in“ incubator maintained at a temperature of 35°C and incubated for 18 - 20 hours. Studies were also made of the flora developing upon I‘11ters incubated upon this medium at MPG. For this pur- pose Pyrex cups were used which were 50 mm. in diameter at the base, 80 mm. in diameter at the top, and 50 mm. deep. \ *Difco Laboratories, Detroit, Mich. 32 Two absorbent pads were placed in the bottom of each cup and moistened with sufficient medium to just cover their surface. After the inoculated membranes had been rolled on in the usual way, the cups were covered with sterile filter paper soaked in sterile water and were sealed with a plastic cover. No sterilization of these covers was found to be necessary, there being no evidence of any contamination. The cups were 1' ineu'bated in a magnetically stirred water bath maintained at a temperature of 4410.5OC. for a period of 18 - 20 hours. The weight of the cups caused the water to come within 2 cm. l, ; or their tops. There was no method whereby the temperature “"4 0f the medium could be determined, but it was assumed to be that: of the bath. At the end of the incubation period colonies showing a metallic sheen were counted using a binocular dissecting Inier'oscope having a magnification of x6.6 diameters. A Source of illumination as nearly vertical as possible was plr'<>V:i.ded as this enabled the presence of the sheen to be recognized more easily . M.P.N. ggtermimtigng. These were used as the stand- and for comparison. Quintuplicate amounts of the water sam- ple were inoculated into 10 ml. of double strength lauryl tt‘those broth and quintuplicate amounts of 1.0 ml. and 0'1 all. into 5 - 6 m1. quantities of the single strength medium. The 0.1 ml. amounts were contained in 1.0 m1. of a U10 Clilution of the sample prepared by pipeting a 10 ml. 33 Portion into 90 m1. sterile 0.85% saline in a dilution bot- tle. This was considered to be more accurate than the method whereby 0.1 ml. of sample is pipeted directly into the medi- um. In addition, the same procedure was carried out using Ma<3<3C>nkey's broth according to British practice. The medium was prepared according to the method described in the Report of the Coliform Sub-Committee (34). In the inoculation procedure, one volume of sample was used for both sets of media, and in order to eliminate external factors as far as possible, one pipet was used for each series of portions pipeted. The same 1/10 dilution of 8amPle was used for the inoculation of 0.1 ml. water into bOth media. In addition, tubes of each medium were inocula- ted alternately. In the case of very heavily polluted waters fur‘ther 10-fold dilutions were prepared and 1.0 m1. portions again used to inoculate tubes of the single strength media. Incubation was in a "walk-in“ incubator maintained at 35°C and tubes were examined after 24 and 1&8 hours. Lauryl tr'yptose broth tubes showing gas production at either of these times were recorded and subcultured, using a triple loop. into tubes of 2% brilliant green bile broth. These t“bee were also examined after 21+ and (+8 hours' incubation at 35°C and those positive, i.e. showing gas production, recorded. Positive tubes of MacConkey broth, recognized by the 31+ presence of both acid and gas, were similarly treated, but a second series of subcultures in the brilliant green bile broth were prepared and incubated in a magnetically stirred water bath maintained at a temperature of hh:0.5°C. Incuba- tion was continued for 1&8 hours at this temperature when ' positive tubes were recorded. From the data so obtained the Most Probable Numbers (M.P.N.) of coli-aerogenes organisms and of E.coli were cal- culated from the tables in the Tenth Edition of Standard Methods (39). Cultural mdigg. In determining the gram staining reaction and IMViC characteristics of pure cultures, the fOIlOTWing procedures were used. Gram's stainzThe Burke and Kopeloff-Beerman modifica- tion (214-) was employed. Indole tests were carried out on 2-day old cultures in be«itto-tryptone broth using Pringheim's'reagent .to detect the presence of indole. For the M.B. and V.P. tests Bacto MB-VP broth was used, the tests being carried out as described in~the RepOPt (34). In order to determine citrate utilization, Bacto 81mmends citrate agar slopes were used. Inoculation was by Stab and any change of color in the indicator after 48 hours noted . 35 Results, Discussion, and Conclusions _A. The comparison offllaurylfi tryptose broth and MacConkey'mfl. The results obtained using both of these media in the 5-tube decimal dilution standard procedure are given in Table I. In all, forty—two samples of water were examined. Identical results in the confirmed test were obtained in the case of sixteen samples which contained no coli-aerogenes bacteria and, in addition, two samples containing these organisms gave the same result. Of the remaining samples, fourteen gave higher M.P.N. detel‘minations when the MacConkey medium was used, and in ten instances the higher count was obtained by the use of laut‘Y]. tryptose broth. The results obtained in all cases are very similar, and if the 95% confidence limits of the- M‘P-N. determinations are used for the comparison, there is no difference between them. In the case of MacConkey's broth, fourteen, or 6.3% of the 226 positive tubes, did not confirm in 2% brilliant gbeen bile broth. With lauryl tryptose, fifteen more tubes were positive initially but of the total of 241 positive tubes, thirty—two, or 13.3%, did not confirm. Thus it ap- Dears that there is very little difference between the two med-la but that MacConkey's broth, in this investigation, tglided to give fewer false positive results, the number of tubes confirmed with both media being almost the same. It .. 1‘ m . ‘ s ... .. . III. [.141 A) ll 36 is considered that the addition of a pH indicator, brom- cresol—purple, to the lauryl tryptose medium would be an advantage in that if tubes showing small amounts of gas with no acid production were considered negative, fewer false positive results would be obtained. In Britain, where pre- sumptive tubes showing acid and gas production in MacConkey's broth are considered to contain coliform organisms without W ‘ subsequent confirmation, this is a decided advantage, and for all practical purposes renders subculture into a confirm- atopy medium unnecessary . i D'Jx1l‘fi mON H N NHN wNN NW1 oo:N TIIIHH Hfl|||||_ oos XHH. eH, H om: NH mH omn HH HH W H own. 0H 0H OHN NH mH . H oHH m m m: N 5 ~ H on m m m: e N H a: i N N mm o o n H S N s _ R s s i H a: s N Hm w m m H . 5H m m :H n o l H M m.o m e N H h m m H w _ W om: NH _ NH om: NH NH w H m OMH m _ 0H omH m OH H H w o o n o o o w 0H W .. “ ooomH m m ooomn HH HH W H W oo0NH mH .oH ooozN NH NH _ H i o. ONm mH mH oowH :H :H _ H m 1, can HH mH . om: NH NH M H w 03H 0H NH ONN HH MH H . mm c m on m m H m: N m an m m H nN m m co m m H mm 0 HH we N N H mN m m mm o w H mH : m mm b m H 0.0 m o mN m m H m.: N N m.s m m H o o m N H H H onEMm poop poo» onEdw umop poop uqu .Hs 00H Lon uoaprnoo o>Huaasmmpm .He 00H non moapHmsoo o>Hpa83mop name aqumwao moqowopom comm um mason m: QH namHnmwao nosomopom filo an an mason m: GH 9. aHHoo .z.m.: 0>HpHmom moss» no .02 nHHoo .z.m.z o> pHmoa mops» mo .0 .o. snosllmuammwmwldwumoq _ suopmlmqummmmummw .ll ‘4 .l Idll‘.| ‘III’IO . I | ‘ ‘3‘ II.“ \‘ u t‘lultllll. 1| 1 .III tti rill-It'll .Ill.‘ ‘uf‘i‘t‘ ll' '1‘..‘|ll - .I .‘u’llil. ‘ ‘ I - -spopm one». we H mama ess-asHeo= m. oxzoomwz 29H: osoHnnooe noHpsHHo onssum HmaHooo on» quae eonHmsno msHSmom .H mqmHpHmoa mops» mo .0: on» song ompmHSOHmo monouOLomuHHoo no empssoo mmHQOHoo oHaemm .Ha 00H Lea wsmHsmwao monmmopomuHHoo 2m: 2m: ooumHSOHmo ,wsHoseoaassoOSm mo .0 .m.z Suopm commit L s.oo:poz :oHpsHHa was»: pameampm 0:» soap ompmHSOHmo mamHsmwpo mosowopooo moHnoHoo wsHosoopmnsmosm no popssz on» mo somHamasoo one .HH MHmma 47 C. _The comparison of the counts of_§:coli obtained at “#00 by the membranefglter method with M-Endo broth MF and by using theg5-tube_d;lutionprocedure with confirmation of tubes positive in the prggumptive test in B.g;§;broth_§£ .3229. In Britain, the enumeration of E.coli type I in water supplies is carried out in many laboratories as a routine procedure. This is accomplished by sub-culturing positive presumptive tubes of MacConkey's medium at 37°C into tubes of either fresh medium or 2% B.G.B. broth, and incubating these in a water bath at 4410.50C.for 48 hours (35). This test has been shown in that country to be practically char- acteristic of E.coli, although infrequently some A.aerogenes strains which can produce gas from lactose at that tempera- ture are encountered. It is apparent that should the membrane filter tech- nique be found adequate and replace the presumptive standard tube dilution procedure, then an alternative method would be needed to estimate the E.coli content of water supplies. This could probably be achieved by picking colonies directly from the filter surface into either MacConkey's medium or 2% B.G.B. broth, followed by incubation at huOC. Such a procedure would, however, be extremely laborious and time- consuming, and it was decided to investigate the possibility of incubating inoculated membranes directly upon the M-Endo broth MF medium at #400. This was accomplished as previously described. The method used is obviously capable of consider- able refinement, but it was felt that if comparative counts #8 could be obtained and the procedure established, then more consideration could be given to the design of suitable ap- paratus. The results obtained are given in Table III. In addition to those presented, a further twenty samples were examined in which no coli-aerogenes organisms could be de- tected. Great difficulty was experienced in finding wells and lakes polluted with fecal material and containing E.coli in detectable numbers. The appearance of the filters when incubated at this temperature was markedly different to those incubated at 35°C. Colonies of sheen-producing organisms were smaller, but the most striking difference was in the numbers of "background“ colonies which had, except in two instances, completely disappeared. Even in these two cases, the num— bers of non-sheen-producing organisms was reduced to negli- gible preportions, and it was possible for large volumes of sample to be filtered before confluent growth was obtained. Nineteen samples of well water were examined which contained coli-aerogenes organisms but in which no E,coli could be detected by either method even though volumes of water considerably in excess of 100 ml. were examined by the membrane filter procedure. In three instances small numbers of E,coli were de- tected by the tube dilution method, and although 200 ml. of water was examined by the membrane filter procedure, no 49 sheen-producing colonies were visible upon the filters after incubation. The remaining ten samples examined gave positive results by both methods. In general the counts obtained by the membrane filter procedure were lower than those calcu- lated from the results of the tube dilution method, but if the 95% confidence limits of the M.P.N. determinations were used in the comparison, only in one sample (3Table III) is the difference significant. .It was decided to carry out cultural studies in order to determine the specificity of both methods for E.coli type I. Colonies were picked as previously described from the filter surfaces and their gram staining and IMViC reac- tions determined. In the case of the tube dilution method, one tube of the highest positive dilution at 44°C was streaked on Levine's E.M.B. agar and again colonies picked and similarly identified. Due to the scarcity of supplies containing E.coli these studies were restricted to seven samples of heavily polluted river water and to one sample from a contaminated well. ' One hundred and forty-six discrete colonies were picked from membrane filter surfaces to lactose broth and all produced acid and gas in that medium showing them to be coliform. Of these, 106 or 72.6% produced gas in 2% B.G.B. broth in 48 hours at 4410.5°C, but only 54 or 50.9% of this 50 total gave the IMViC reactions typical of E.coli type I, the remaining 49.1% being identical to A.aerogenes except for the ability to produce gas at 4u°c. From the tubes of MacConkey's broth confirming at 44°C, only cultures of E.coli type I were obtained. From the results obtained with the membrane filter procedure it might have been expected that orgamisms of the A.aerogene§ 44°C positive type would beqfound. Their absence may be due to the fact that only a small number of tubes were streaked, and these were of the highest positive dilution. In addi- tion, the E.coli tended to predominate and would probably outgrow the A.aerogenes if the two were present together. A number of isolates from the filters representing 27.4% of the total, did not produce gas when incubated in. B.G.B. broth at 44°C for 48 hours and with the exception of two cultures (44--) corresponded to the IMViC classifica- tion of A.aerogenes. Thus it appears that a high proportion of coliform organisms can grow upon this medium under the conditions of incubation described which are not true 44°C positive cul- tures, and this discrepancy, if found to hold good for all types of water supplies, is considered to be too large for the adoption of this procedure for routine examinations. Again, as has already been pointed out, false negative results are obtained and in any event the performance of this medium at 37°C is such that there is considerable doubt as to its usefulness in the bacteriological examination of water. 51 In conclusion, it may be said that the procedure is not satisfactory, the counts obtained being much lower than those calculated from the results of the standard British procedure for the enumeration of E.coli type I in water supplies. This discrepancy is no doubt due to the combined toxic effects of the indicator system, inhibitory agents, and initial high incubation temperature used. It is greater than appears at first sight because it has been shown that of the sheen-producing colonies counted only 72% are capable of producing gas from lactose at 44°C, the remainder being coliform organisms other than E.coli type I. It is consid- ered that these discrepancies are such as to preclude the use of this procedure for the routine enumeration of E.coli type I in water supplies. 52 . , .o m pm mpmou o>HpQESw0LQ o>Hproq an ompmhpmnoaop mm mEmHnmmpo mosmmopmmuHHoo pmsHmusOo mmH smm oso N ooo.mH oom.m m m oom.HH omH.m on em as oo om Ha a: msa m m * m.o mm m : m.o .Ha oom :H anmmpo o A o m.s .Hs oom 2H scoops o A o m.: .H& oom 2H psmmpa o o so .Ha oom nH usmmn< . _ o so .Hu oom :H pcmmp< - o so .Ha oom sH pecans . o so .Ha oom 2H psompo u o so .Ha oon :H psmmpa o o so .Ha ooo :H pnmmna o o so .Hs oos nH pnmmno o o so .Ha oos :H pammna o o so .Hs oos eH pnmmp< o o so .Ha oom :H psomp< o o so .Ha oom qH oceans o o so .Ha oom :H ansmna o o so .Ha oom :H pzmmno o o so .Ha oom :H pqmmpa o o so .Ha ooN :H pcomnm o o so .Hs oom nH scoops o o so .Ha oom nH pnmmn< o o so .Ha oom sH pnmmn< o o so In! 1: .Ha oom nH onompa [m o oo:: so spopn mom mm :H ooapHe onamm .Hs ooH“ m 4 1:00 mecosummnSm soHns comm um npopn pom mBmHsmeo opmoH on «$83832 :H 3338 mops» co $956 on» mmmownmmwwmwflw namswswwmsmwmmewmewL on m m no mo m mamm . a hog H 00. m: I a o o H H H H ooH H m z .m.z spasm comma: cos: pa noHuoapHcsoo ssz oospo: noHusHHn ops»- was gone omomH30Hmo HHoo.m so sonssz as» new msoHnSooe .m.z on» suHmp cos: as ooumnsqu m: spopm OUHWOE HMO wfidQOHOPOQ EDHGOHOO deONHUOleQOOSW .MO #09332 0.89 MO SOMHLMQEOU mg sHHH Edam... 53 The Present Status of the Membrane Filter Procedure in the Bacteriological Examination of water In Britain at the present time the membrane filter procedure has no official standing in the enumeration of coliform organisms in water and is not recommended for this purpose. In America,where nearly all of the work done in con- nection with the development of this system of analysis for the bacteriological examination of water has been carried out,the mechanics of the filtration procedure have become well established and are regarded as being satisfactory. As has been indicated, a great variety of media have been de- velOped and used for the growth, recognition, and enumeration of the coliform group of organisms. By far the most atten-” tion has been given to modifications of the Endo type of medium which utilizes basic fuchsin and sodium sulphite as an indicator system for the recognition of coliform colpnies. It has become apparent that this indicator is toxic to at- tenuated coliform cells and is not by itself a suitable in- hibitor of non-coliform organisms. Other materials such as brilliant green, bile salts, sodium desoxycholate, and sodium lauryl sulphate have been included in most formulae for this purpose; and, in an effort to overcome their toxicity to coliform organisms, complex sources of vitamins, amino-acids, and micronutrients have been supplied. This type of medium would appear to have reached the peak of its development in 54 the M-Endo MP broth described and which has been shown to be unsatisfactory for-the examination of heavily polluted nat- ural water supplies. In the Tenth Edition of Standard Methods for the Examination of Water, Sewage and Industrial Wastea,a filtra- tion procedure and a medium (EHC Endo broth) are described for use in the enumeration of coliform bacteria in water. It is unfortunate that the conditions under which it may or may not be used are not clearly defined since while it is stated that "It must be understood that this (the membrane filter procedure) is in no way a standard technique, and it cannot be considered an acceptable substitute for the tube dilution method,” the next sentence reads, "When the limita- tions of the test are fully recognized and the difficulties of the interpretation of the results are known, the tech- nique may be used." In a more recent memorandum the subcommittee dealing with this section have attempted to clarify the position. The advantages of the procedure are described and the limita- tions discussed. The subcommittee in conclusion considers that "For the applications discussed above (the testing of water potability) and only when used with full recognition of the limitations noted, the membrane filter technique as given in the Tenth Edition of Standard Methods, or its demon- strated equivalent will yield results comparable to those of the Standard Completed Test for determining the presence of 55 coliform bacteria in potable waters." In addition, it is also stated that “In no instance should the membrane filter method be adapted for testing water potability until adequate parallel testing of duplicate samples by the membrane filter technique in comparison with the standard completed test has demonstrated beyond reasonable doubt the relationship of the results obtained by the two methods." It is apparent that the official status of the proce- dure is by no means clearly defined. It is felt that the obvious fundamental advantages inherent in the procedure have been so great as to obscure the fact that without the devel- opment of a suitable medium for the growth, recognition, and identification of coli-aerogenes organisms, the results ob- tained with it are valueless. The position is made more serious by the fact that the manufacturers of the necessary equipment and media have given the method a great deal of publicity, and media are now avail- able and are being sold for use in this respect which are not supported by any published data. A new word has been coined for their application, the label on one bottle indicating that the contents are ”for use in the detection and quanti- tation of coliform organisms.‘I Presumably this is not the same thing as quantitative estimation. It is the Opinion of the author that none of the media develOped so far are suitable for use with the membrane fil- ter procedure for the enumeration of coliform organisms in 56 all types of water supplies. It is felt that a different approach is required to the problem and that a re-examination of the growth requirements of the coli-aerogenes group, to- gether with information concerning the biology of the organ- isms associated with them in water supplies, is required,and may lead to the deve10pment of a different type of selective medium based rather upon nutritional requirements than upon the use of inhibitory materials. REFERENCES (1) Childs, E., and Allen, L. A., 1953, “Improved Methods for Determining the Most Probable Number of Bacterigmcoli and of Streptococcus Faecaligs” J. Hyg. 51, No. #, 468. (2) Clark, H. F., Jeter, H. L., Geldreich, E. E., and Kabler, P. W., 1951, “The Membrane Filter in Sanitary Bacteriology." Pub. Hlth. Rep. fig, 951- (3) Clark, H. F., and Kabler, P. W., 1952, ”The Membrane Filter in Water Quality Tests.” Amer. J. Pub. H1th., 32, 385. (a) DeBord, G. G., 1917, "The Fuchsin-aldehyde Reaction on the Endo Medium." J. Bact., g, 309. (5) Endo, 5., 1904, 'Ueber ein verfahren zum Nachweis der Typhus Bacillen.‘I Centralblatt fur Bakteri- ologie, 0rig.,‘35, 109. (6) Fifield, C. W., Hoff, J. E., and Proctor, B. E., 1956, “A Method for the Enumeration of Organisms of the Coliform Group in Pasteurized Milk Using the Millipore Filter.“ In Press. (7) Floyd, T. Mg, and Dack, G. N., 1939, J. Inf. Dis., §3, 29. (8) Gassner, G., 1918, ”Metachromgelb als Hemmungsmittel fuer Kokken und Sporenbildner und seine Verwendbarkeit fuer Naehrboeden zur Typhus und Bhurdiagnose.“ Zentre. Bakt. Parasitenk. Abt. I 0rig., 39, 120. (9) Goetz, A., and Tsuneishi, N., 1951, ”Application of Molecular Filter Membranes to the Bacterio- logical Analysis of Water." J. Amer. Water Works Assoc., flfiJ 9H3. (10) Goetz, A., Gilman, R. H., and Baun, A. M., 1952, "Application of Molecular Filter Membranes to Specific Problems in Water Analysis." J. Amer. Water Works Assoc., 33, #71. 57 (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) 58 Hajna, A. A., and Damon, S. R., 1959, ”Coliform Deter- mination in Water by a Single Step Technique Using the Membrane Filter.” Pub. Hlth. Reports, 62, 58. Holt-Harris, J. E., and Teague, 0., 1916, "A New Cul- ture Medium for the Isolation of Bacillus Typhosus from Stools.“ J. Infect. 018., E. 596. Howard, N. J., and Thompson, R. E., 1926, "Isolation of the Colon Group in blater.‘I The Canadian Engineer, £8, #13. Jeter, H. L., Geldreich, E. E., and Clark, H. P., 1959, "Type Differentiation of Coliform Organisms with Membrane Filter Technique.“ J. Amer. Water Works Assoc., 26, 386. Water Works Kabler, P. w., 1951, Discussion. J. Amer. Assoc., £3, 969. Kabler, P., 1954, ”Water Examinations by Membrane Filter and Most Probable Number Procedures.” Am. J. Pub. Hlth., fig, 379. Kabler, P. W., and Clark, H. F., 1952, "The Use of Differential Media with the Membrane Filter." Amer. J. Pub. H1th.,‘&g, 390. Levin, G. V., and Laubausch, E. L., 1959, "Membrane Filter and Emergency Water Supply Control." Amer. J. Pub. H1th., 35, 55. Levine, M., 1918, "Differentiation of B.coli and B.aerogenes on a Simplified Eosin-Methylene Blue Agar.” J. Infect. Dis., 23, #3. Levine, N., 1921, "Bacteria fermenting lactose and their significance in water analysis.“ Iowa State College of Agriculture and Mechanic Arts, Official Publication, 29, No. 31, Bulletin 62. Levine, M., 1993, "The effect of the concentration of dyes on differentiation of bacteria on Eosin- methylene blue agar.” J. Bact., 35, #71. MacConkey, A., 1900, "Experiments on the Differentia- tion and Isolation from Mixtures of the Bacillus Typhosus by the Use of Sugars and the Salts of Bile.” The Thompson-Yates Laboratories Report, 3, #1. (23) (24) (25) (26) (27) (28) (29) (30) (31) (32) (33) 59 Mallmann, W. L., and Darby, C. N., 1941, “Uses of a Lauryl Sulfate Tryptose Broth for the Detec- tion of Coliform Organisms.” Amer. J. Pub. Hlth., 3;, 127. "Manual of Methods for Pure Culture Study of Bacteria,” 1946. Society of American Bacteriologists, Biotech Publications, Geneva, New York. Margolena, L. A., and Hansen, P. A., 1933, “The Nature of the Reaction of the Colon Organism on Endo's Medium.” Stain Tech., 8, 131. McCarthy, A. M., 1956, "Studies on the Membrane Filter Technique and its Application to the Detec- tion of Coliform Organisms in Water.“ Thesis for the Degree of Ph.D., Michigan State Uni- versity. McCarthy, J. A., 1955, ”Comparative Coliform Densities in Water by the Membrane Filter Test.“ Amer. J. Pub. Hlth., 45, 1569. McKee, J. E., Derby, R. L., Goetz, A., Noble, R. E., and Streicher, L., 1953, "Technique for the Bacterial Examination of Water with Molecular Filter Membranes." J. Amer. Water Works Assoc., 35, 11. Millipore Filter Corporation, Watertown, Mass., 1957, Personal Communication. Neuberg, C., and Nord, F. F., 1919, ”Anwendungen der Abfangmethode auf die Bakteriengarungen :1 Acetaldehyd als Zwischenstufe bei der Vergarung von Zucker, Mannit and Glycerin durch Bacterium coli, durch Erreger der Rhur und des Gasbrandes.” Biochemische Zeitschrift, 26, 13. Neumann, H. G., 1951, Discussion, J. Amer. Water Works Assoc., £3, 975. Prescott, S. C., Winslow, C. E. A., and McCrady, M. H., 1945, "Water Bacteriology." 6th ed., Wiley and Sons, Inc., New York. Presnell, M. W., Arcisz, W., and Kelly, J. B., 1954, "Comparison of the M.F. and M.P.N. Techniques in Examining Sea Water.” Pub. Hlth. Rep. Q9, 300. (34) (35) (36) (37) (38) (39) (40) (41) (42) (43) (44) (45) 60 Report of the Coliform sub-committee, 1949, “The Classification of the Coli-Aerogenes Bacteria." Proc. Soc. Appl. Bact. g, 3. "Reports on Public Health and Medical Subjects - The Bacteriological Examination of Water Supplies," 1939. H.M.S.O., London. Schutz, F., and Kruse, H., 1950, “The Use of Membrane Filters in the Bacteriological Examination of Water, More Especially for the Quantitative Examination of Coliform Organisms.I Bull. Hyg. Lond., 23, 289. Slanetz, L. W., and Bartley, C. H., 1955, “Evaluation of Membrane Filters for the Determination of Coliform Bacteria in Water." Appl. Microbiol., 2. 6- Standard Methods For the Examination of Water and Sewage, 1923. Fifth Edition, American Public Health Association, New York. Standard Methods for the Examination of Water, Sewage, and Industrial Wastes. Tenth Edition, American Public Health Association, New York. Streicher, L., 1951, Discussion. J. Water Works Assoc., 33, 975- Amer. 'Supplemental Technical Information Regarding Bacterio- logical and Other Analytical Procedures," 1955, Millipore Filter Corporation, Watertown, Mass. Taylor, E. W., 1953, “Use of the Membrane Filter in the Bacteriological Examination of Water." Chem. and Ind., 1, 124. Taylor, E. W., Burman, N. P., and Oliver, C. W., 1953, "Use of the Membrane Filter in the Bacterio- logical Examination of Water.“ J. Applied Chem., 3, 223. Taylor, E. W., Burman, N. P., and Oliver, C. W., 1955, "Membrane Filtration Technique Applied to the Routine Bacteriological Examination of Water." J. Inst. water Engr., 2, 248. Thomas, H. A., and Woodward, R. L., 1955, “Examination of Coliform Density by the Membrane Filter and Fermentation Thbe Methods.“ Amer. J. Pub. Hlth., 43, 11. (46) (47) (48) (49) (50) (52) 61 Thomas, H. A., Woodward, R. L., and Kabler, P. W., 1955, “Water Potability Control with Membrane Fil- ters.” Paper presented at the 44th meeting of the Committee on Sanitary Engineering and Environment, National Research Council, Division of Medical Sciences, Washington, D. C., Oct. 7, 1955. U.S.P.H.S. Environmental Health Service, Reports, No. 14, Oct. - Dec. 1952, Activity U.S.P.H.S., U. 8. Dept. of Health Education and Welfare, 1953, Environmental Health Center Activity Report,1__6_,13. Velz, C. J., 1951, "Graphical approach to statistics IV, Evaluation of Bacterial Density." Water and Sewage Works, 8, 66. Woodward, R. L., 1952, "Results of Statistical Analysis of Data Comparing the Dilution Method and Membrane Filter Method for Estimating Coli- form Densities in Various Waters.” Environ- mental Health Center Beport No. 11. E. 3., Bode, L. J., and Hayward, A. E., 1942, "Mechanism of the Selective Action of Eosin- Methylene Blue Agar on the Enteric Group.” Stain Tech., 31, 11. Wynne, Yee, G. S., Krabek, W. E., and Schaufus, C. P., 1953, “New Medium for Bacteriological Analysis With Molecular Filter Membranes.” J. Amer. Water Works Assoc., 43, 945. MS 1957. MPH Studies with media for use with membrane filter in ‘ Panes, John J ‘ enumeration of coliform Panes, John J ."S 1957 WE Studies with media for use with membrane filter in enumeration of coliform organisms in water. MICHIGAN STATE UNIVERSITY LIBRARIES I III“ IIIIIIII I 19 3 1213 03142 7