‘1... '..;/‘ STQSEES (3%? THE USE €35: ESE-6Y3 "iSCLE‘? AS AN. .ERSHSBETC‘IRY .QGERT M A CGN‘HRMAHON MEQEUM FQR T”HE DE'E'ECWCN OF CQLSFGRM ORGANSSMS 'fhesés for H10 Dagmar 02‘ M. S. [\MCHIGAN STATE" CGLLEGE Hederick Jusi‘ R35? 1953 “-«~A I m— r‘fi‘TT—- 1‘” r.__ This is to certifg that the thesis entitled fl 4'1 'l . M ‘ - "l 1'" 5 V " ' ~ u l ~J. - l _ L. _.. _ .. thl .9“: -- - v,” - -- t “rt ‘ s. , . 13”” . _, I A -\/ . *.L U4 \ fl .5 fl ‘4 (5 'V '5 ' ’. (a A fl ,- 5 v V‘ LIT . "f u 4* -——> \Jl . x -A— L) has been accepted towards fulfillment of the requirements for l. 'd.“ ." Lem degree in S C L "_‘."‘.<_"' e Maj r professor ‘Y fi'rlv."\‘l‘l1 ”K “(VI T l)ate ’3 " 'l " “'1 l ’ 1 (A. 0169 -u' ‘3 _______l;_k‘_ 13.4.3.- STUDIES ON THE USE OF EHHYL VIOLET AS AN ILEIBITORY AGEJT IN A CONFIRMATION HEDIUM FOR.TLE DETECTION OF COLIFORM ORGANISHS BY Frederick Just Post A THESIS Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Bacteriology and Public Health 1953 I //' I I 3“} ’/ Ila-J \rg, ACKNOWLEDGMENTS The author wishes to express his appreciation to Dr. W. L. Mallmann, Professor of Bacteriology and Public Health for his able guidance and advice, and to Mr. O. E. McGuire of the Michigan Department of Health for his cooperation in the field studies. ‘2£“?=:!¢1 TABLE OF COHTEETS Page FTHOBUCTION . . . . . . . . . . . . . . . . . . . . . . . . 1 Purpose of the study . . . . . . . . . . . . . . . . . 2 REV 3W OE LITERATURE . . . . . . . . . . . . . . . . . . . . 3 KATERIALS AND KLTEODS . . . . . . . . . . . . . . . . . . . 8 Cultures . . . . . . . . . . . . . . . . . . . . . . . 8 Culture media . . . . . . . . . . . . . . . . . . . . . 8 Growth studies . . . . . . . . . . . . . . . . . . . . 9 Field comparison . . . . . . . . . . . . . . . . . . . 10 RESULTS AID DISCUSSION . . . . . . . . . . . . . . . . . . . 11 Growth studies . . . . . . . . . . . . . . . . . . . . 11 Comparison with brilliant green bile . . . . . . . . . 22 SUYXARY AJD COECLUSICNS . . . . . . . . . . . . . . . . . . 37 L I ff ~1ka 'LrU-wi-IE C I jED . O C O O O O O O O O C O O O O O I O O C O 39 Table II. III. IV. V. VI. VII. VIII. IX. X. XI. XII. LIST OF TABLES Title Pepulations of E, coli in varying concentrations of p. ethyl violet n lactose broth . . . . . . . . . . . Populations of E, 3211.1n varying concentrations of ethyl violet in tryptose lactose broth . . . . . . Populations of E, coli in varying concentrations of ethyl violet in lauryl tryptose broth . . . . . . . Source and number of samples . . . . . . . . . . . . The comparative confirmation of positive presumptive tubes by the use of brilliant green bile broth and ethyl violet tryptose broth . . . . . . . . . . . . Time of confirmation related to time of becoming pre- sumptive positive among all samples . . . . . . . . Distribution of Most Probable Humber (HEN) divergency among the samples . . . . . . . . . . . . . . . . . Time distribution of confirmations among the samples not agreeing in MPH . . . . . . . . . . . . . . . . Distribution of tube differences among the non-agree- ing samples . . . . . . . . . . . . . . . . . . . . Results of completed test on samples showing difference in MPN from parallel tubes . . . . . . . . . . . . Occurrence of false tests as determined from Table X. Number of positive confirmatory tubes in each medium by sample, and the medium showing the correct results by the completed test . . . . . . . . . . . . . . . Page 13 15 17 24 2M 26 2’7 28 30 33 3h 36 LIST OF FIGURES F IGURE Title Page 1. Effect of various concentrations of ethyl violet on the growth curves of g, ggli_in lactose broth . . . 1M 2. Effect of various concentrations of ethyl violet on the growth curves of E, ggli_using tryptose lactose broth . . . . . . . . . . . . . . . . . . . . . . . 16 3. Effect of various concentrations of ethyl violet on the growth curves of g, ggli_using lauryl tryptose broth . . . . . . . . . . . . . . . . . . . . . . . 18 h. Effect of autoclaving ethyl violet tryptose lactose broth 0 o o o o o o o o o o o . o o o o o o o o o o 21 INTRODUCTION Since the early beginnings of water bacteriology attempts have been made to improve and make more exact the methods of determining the sanitary quality of a water supply. Diseases, which are frequently transmitted through polluted waters, have been successfully controlled to the point where they are no longer considered a menace. Much of the credit for this has been due to the constant effort of bacteriologists to make more stringent the criteria and limits of safety of pollution, especially through the agency of the coliform test. Standard Methods for the Examination of Water and Sewage (19N6), which is the accepted standard for bacteriological procedures in water analysis, defines coliform organisms as;“all aerobic and facultative anaerobic, gram negative, non-sparing bacilli which ferment lactose with gas formation." One of the permissible media allowed by the above "Standard Methods" for confirming the presence of these coliforms is brilliant green lactose bile broth, which is considered more toxic for gram positive organisms than for gram negative organisms. Many authors (Ruchoft, 1926; Hale, 1927; Salle, 1929; Mallmann and Hepler, 1936 and others) have found that brilliant green bile is toxic to many coliforms to the extent of completely inhibiting the weaker strains, lengthening the generation times or inhibiting the lactose fermenting powers of many others. There is, then, a constant search for a better medium to take its place; one that is less toxic and more nearly approaches optimum conditions for the growth and multiplication of coliform organisms without removing its selectivity. Following the earlier work of Jolliffe (1948), Bordt (1951), and Litsky, Mallmann and Fifield (1952), this dissertation concerns itself with the attempt to incorporate an aniline dye, ethyl violet (hexaethyl triamino triphenyl methane), in an enrichment medium.to be used in con- firming the presence of coliform organisms. Purpose 93.333.533gyg The purpose of this study was twofold: first, to develop the growth curves of a strain of Escherichia coli in various media containing varying amounts of ethyl violet dye in order to determine the most efficient combination; and second, to compare one dye—medium combination with brilliant green bile broth in the "Standard Methods" examination of various Michigan waters. The test organism was grown in three media; "Standard Methods" lactose broth, tryptose lactose broth, and lauryl tryptose broth using concentrations of ethyl violet ranging from l-lOT to l-SOOT (T = thous- and). The growth cycle study was limited to the lag and early log phases which were considered more important in this type of study than the total number of organisms at 2h-hours. The comparison with brilliant green bile was carried out according to Standard Methods for the Examination of water and Sewage (19Mb), at the Michigan Department of Health Laboratories, by inoculating dupli- cate tubes. REVIEW OF LITERATURE In reviewing the literature on growth curves of E: 9211, little was found relating directly to the subject. Salter (1919), made some observations on the rate of growth of B, ggli_in lactose broth con_ taining various concentrations of crystal violet and brilliant green dyes. He established the effect of the concentration of the dye by its effect on the generation time, or the rate of.growth, for which he utilized the law of geometric proportions. He found that crystal violet decreased the growth rate at l-lM (M== million) and its greatest effect was in the lag phase. Brilliant green did not inhibit at l-bM but completely inhibited at 1-600T. Mallmann and Darby (1939), utilized a technique using a minimal inoculum and subsequent sampling every two hours in order to determine the effect of various factors on the lag and early log phases. In this study they also developed a tryptose lactose broth medium.that was far superior to lactose broth as an enrichment medium. Later work by Litsky, Mallmann and Fifield (1952), using the above technique and medium, found that crystal violet,and brilliant green were quite toxic to p, 22;}, the latter being more toxic at l-lM, al- lowing an increase to 280 organisms, while crystal violet in the same concentration allowed 270T organisms in the same time period from the same minimal inoculum. Prior to 1920 attempts at confirming the presence of coliform organisms were limited to the use of solid media while experiments proceeded using dyes in the presumptive enrichment media. Meur and Harris (1920), first suggested the use of a lactose-bile-brilliant green medium for the presumptive test and Winslow and Dolloff (1922) determined the limiting concentrations of the bile and brilliant green. Because of many conflicting reports on the results of using this broth, the American water Works Association, Standard Methods of water Analysis, Committee No. 1 studied this medium intensively and found it less suitable as a presumptive enrichment medium.than lactose broth.' Work by Jordan (1927) and others with optimum bile-brilliant green ratios, showed that it compared unfavorably with lactose broth as a presumptive medium. In this study, however, Jordan used the brilliant green bile broth as a confirmatory medium and found it very satisfactory as such. These results and others led toithe inclusion of this medium in the 1933 "Standard Methods", and favorable results were later re— ported by France (1936), and Mallmann and Hepler (193b), although the latter noticed inhibition to some coliforms and the failure to suppress some spore formers both aerobic and anaerobic. Again in 19Ml, Mailmann and Darby demonstrated that some organisms, when placed on eosin methylene blue agar or in brilliant green bile broth, lost their ability to ferment lactose or else were inhibited from growing. Other liquid media, previously proposed as presumptive media, were examined for the possibility of obtaining a better confirmatory medium, just as occurred with brilliant green bile broth. Ruchhoft and Norton (1935), compared lactose broth, and lactose broth followed by brilliant green 2 percent bile broth, formats ricinoleate broth, and MacConkey broth as confirmatory media. Carrying all tubes through the completed test, he found that the "Standard.Methods" brilliant green bile medium compared unfavorably in that more coliform isolates were obtained from all of the confirmatory methods together than from any one alone. Most other confirmatory media have eventually been discarded because of toxicity, non—selectivity, or that theyvere not sufficiently sensitive for use. McCrady (1939), in comparing MacConkey's broth and "Standard Methods" lactose broth for the presumptive test, found that lactose broth followed by brilliant green bile yielded.the best results al- though brilliant green bile gave a few false positive tubes. Richey (19Ml), and Wattie (19u3), found brilliant green bile, on the whole, satisfactory, but Shane (19u7), found that, by using lauryl sulfate broth as a presumptive medium, many of the false positive tubes occurring in the brilliant green bile confirmatory medium were eliminated. A more complete and detailed evolution of sanitary water bacteri- ology and the methods and.media used can be found in the book "Water Bacteriology" by Prescott, Winslow, and McCrady, (l9uo). The first mention of ethyl violet as a bacteriostat found in the literature was in a paper by Petroff and Gump (1935), who used it as one of a series of 130 dyes and allied compounds tested for bacterio- static action against several gram positive and negative organisms. This study utilized the method developed by Churchmann (1912), in his studies on gentian violet. The results were apparently disappointing to the authors who found that ethyl violet, among others, was less toxic to gram.positives than gentian violet but more than brilliant green. It is assumed that ethyl violet was not toxic to the gram nega- tive organisms studied for it was not included among those compounds active against this group. Brilliant green was included and was about equally toxic for the gram negative organisms (including 3, 991;) as for the gram positives. It is interesting to note that these authors prepared the next higher homolog of ethyl violet, npproply violet, and found it less toxic to gram positives and apparently nonptoxic to gram negatives. The above authors, however, did not appear to recognize the pos- sibilities of ethyl violet for not until the work of Darby (1943), was this dye mentioned again. Using agar and broth containing ethyl violet, brilliant green and several other agents to determine bacteriostasis, he found that ethyl violet in agar showed marked inhibition to gram positives and very little to gram negatives. In tryptose broth the bacteriostatic titer was considerably higher. Using a tryptose broth base and comparing the growth cruves of E, ggli_with brilliant green bile broth, he found that the broths compared favorably when a l-200T concentration of ethyl violet was used. This combination, however, showed slight toxicity to coliforms as did the brilliant green bile medium. He also found that ethyl violet was not 100 percent effective against gram positives. i In studying the growth curves of several enteric organisms, using a base medium of lauryl sulfate broth, Jolliffe (19H8), found that ethyl 7 violet was less toxic than brilliant green or crystal violet but that all three were about equal with respect to inhibition of several gram positive organisms. Bordt (1951), found that ethyl violet in a concentration of 1-80T in tryptose lactose broth inhibited the gas production of some coli- forms and when the concentration reached l-333T no inhibition was ex— hibited. His studies also indicated that this dye was more toxic to E, coli than to either Aerobecter aerogenes or E, freundii. Litsky, Mallmann and Fifield (1952) found that ethyl violet showed markedly less toxicity to g. 0011 at l-ZOOT concentration than either brilliant green at l-lM or crystal violet at l-lM and that all three were inhibitory to Bacillus subtilis from l-lM to l-lOM concentrations. HATERIALS AND METHODS Cultures. Escherichia coli, Cinncinnati strain 198 was obtained from D. Muentener of the Michigan State College Department of Bacteri- ology and was maintained and transferred daily on tryptose glucose ex- tract agar (TGE) and incubated at 35 - 37C. The IMViC formula was the usual Eh 92l$.+'+""' This organism was used for all of the growth curve determinations. Culture media. Three of the media used; brilliant green bile broth, lactose broth and lauryl sulfate broth, were prepared from the standard dehydrated Difco products and all media used were autoclaved at 15 pounds for 15 minutes. The fourth medium, tryptose lactose broth, was prepared according to the formula suggested by Darby and Hallmann (1939). and is as follows: Bacto-tryptose 2 $ Lactose 0.5% KHzrou 0.15% NaCl 0.5% pH before sterilization 6.8 The description of the dye used in this study is as follows: Ethyl Violet 63, Lot no. 12552, CI*b82, 57.5 percent dye content, manufactured by the National Aniline and Chemical Company. A stock solution of the dye was prepared by dissolving 0.17391 gms in N cc. of ethyl alcohol and making up to 100 ml. Ibis gave a working solution of 1-1000 actual dye content. Sterility was determined by plating in TGE and incubating at 35-370- The media used in the tests were prepared and added to 250 m1 flasks in such a manner that after autoclaving and addition of the pre- determined volume of the dye solution, the total volume was 99 ml. In all the preliminary growth curve studies the dye concentrate was added aseptically to the desired medium after autoclaving and immediately before inoculation. Near the end of these experiments it was found de- sirable to determine the effects of autoclaving on the dye medium com- bination in which case the dye concentrate was added prior to steriliz- ing. During the comparison with "Standard Methods" brilliant green bile in the field, the dye was added to the medium, tubed in fermentation tubes, plugged, and autoclaved at 15 pounds for 15 minutes. Growth studies. Cultures were maintained on TGE agar slants by transferring daily for at least five days before each trial. Stock cultures were refrigerated at NC and transferred monthly. At the time of inoculation of the flasks, the slant containing the culture was washed with physiological saline and adjusted to a known density by visual comparison with a McFarland turbidity standard. Appropriate dilutions were made in physiological saline so that after the addition of 1 ml to the medium, each flask contained between 20 and 80 organisms per m1. One ml samples were removed at time intervals of two hours, begin— ning immediately after inoculation and ending at six hours. The flasks were vigorously swirled, alternately clockwise and counterclockwise at least 80 times before sampling. Using the appropriate dilutions, the samples were plated in duplicate using TGE as the plating medium. The 10 flasks and plates were incubated at 35 - 37 C as recommended by Boniece and Mallmann (1950), and counted at the end of 29 hours using a Quebec colony counter. Eield comparison. This portion of the study was undertaken with the cooperation and at the laboratories of the State of Michigan Depart- ment of Health. It consisted of using the ethyl violet medium.as a confirmatory test for the presence of coliforms by using it in parallel with the brilliant green bile medium test on various Michigan waters. Transplants were made from tubes of standard lactose broth showing gas to the brilliant green bile medium and the ethyl violet medium by pipetting 0.1 ml into each and incubating at 37 C for #8 hours. Tubes showing gas in both brilliant green bile and ethyl violet were recorded and discarded. Parallel tubes showing different results at the end of #8 hours were examined in the attempt to recover coliforms, from one or both, by using the "Standard Methods" completed test. This test con— sists of streaking on eosin methylene blue agar plates and incubating for 2% hours. Isolated colonies are inoculated into standard lactose broth and streaked on agar slants. Gram stains are made from the agar slant at 2% hours and examined if the lactose broth tube shows gas at 2” or 48 hours. If spores are present further isolation is undertaken and the completed test repeated again. ,ESULTS AFB DISCUSSION All the growth curves in this study were determined in exactly the same manner; removing one ml samples, diluting and plating in duplicate. This technique was adapted from.the work of Darby and Mallmann (1939), who did extensive work with growth curves in developing tryptose lac- tose broth and lauryl sulfate tryptose broth. These authors thought that studies on bacteriostats and inhibitory media should be made dur. ing the lag and early log phases, when the young cell, which has been placed in a new environment and is commencing to divide, is very sus— ceptible to adverse conditions. Itwes assumed that the earlier the lag phase can be overcome the more efficient the medium in isolation work. In view of this concept the growth curve was studied only for the first six hours since the lag phase should be of considerably shorter dura- tion to produce good results. Escherichia coli was the organism of choice for these growth curves since Bordt showed evidence that ethyl violet was more toxic to this coliform than to g, aerogenes or the Intermediates. The results of the growth determinations are given in Tables I, II, and III and Figures 1, 2, 3, and M. A.total of eight different dye con- centrations was added to the three media under test in order to deter- mine the relative effect of the concentration. For each group of flasks a control, containing no dye, was used to determine the uninhibited growth in that medium; these are included in each graph in order to make a comparison with the dye curve. 12 The first medium to be tested was "Standard hethods" lactose broth, (Table I and Figure 1), using concentrations of dye l-lOT, l-5OT, l-lOOT, l-2OOT, l-UOOT, and l-SOOT. An examination of Figure 1 shows that dye concentrations less than l-lOOT were extremely toxic to E, 99;; and l—lOOT still someWhat so. The slope of the growth curve does not approach that of the control until the dye concentration is l-4OOT and here the lag phase shows a greater decline of numbers in the first two hours than the control due to the death of susceptible organisms, presumably at the time of dividing. At a dye concentration of l-SOOT the two curves are very nearly alike. It is probable then, from these results, that the 1-UOOT concentration represents the maximum concentration that could be used, and since the initial lag phase drops slightly it would probably be less toxic if a concentration less than l-HOOT were used. Tryptose lactose broth was the next medium to be tested for its suitability as an ethyl violet base, Table II, and Figure 2. The first noticable difference in results from the lactose broth is that any particular dye concentration is less toxic in this medium. A concen- tration of l-lOT roughly corresponds to l-SOT in lactose broth and this similarity holds for each one tested. The results indicate that l-200T is the maximum concentration that can be used, for there is no appreci- able loss of organisms in the first two hours although the generation time in the lag phase is slower than in the control. The slope of the l-HOOT concentration very nearly approaches that of the control so that the safest concentration would very probably lie about l-300T since here one would expect a lag phase generation time only slightly slower than the control and the log phase rate about the same as that of the control. TABLE I Populations of E, coli in varying concentrations of ethyl violet in lactose broth humber of bacteria per ml. Concentration of dye 13 Hrs 1-10? 1-50? l-lOOT l-2OOT 1;h00T l-SOOT Control 0 l“ 33 43 52 72 59 78 H 2 1 25 23 33 36 98 200 a“: , g u 0 16 11+ 160 720 2500 7300 6 0' 6 50 1000 6000 €21,000 M10,000 0 35* 167* 42 52 56 47 56 N 2 1 SO 31; 42 31 56 77 ale 5 u 0 40 15 79 590 1600 3800 a: 6 0 27 30 730 6500 35,000 220,000 0 1+7 51 51 49 40 m 2 3s '40 45 no 52 a": g 4 31 89 530 550 2400 a: 6 21 510 7800 33,000 240,000 *Taken from runs with excessive counts. Howeven it does show the trend. Figure 1 Effect of various concentrations of ethyl violet on the growth curves of E. coli, in lactose broth'. l-lOT dye 1-50r dye Log of cells per ml. Time in hours *Typical curves of at least 3 individual runs in TABLE II Populations of Eh coli in varying concentrations of ethyl violet in tryptose lactose broth. “ “WV Number of bacteria per ml. Concentration of dye u.“ 15 Hrs 1-101" 1.50: 1171001? l-BOOT w1'.-1~00T l-SOOT Control H, 0 26 39 35 1+6 36 45 47 a 2 15 31 26 1+3 be 59 85 2 1t 10 25 68 330 1550 2200 M700 6 8 55 290 (7000)* 60,000 85,000 260,000 0 32 1m 51+ 59 61 56 58 N 2 15 37 I42 50 71 75 110 "a“ u 12 33 76 390 11150 2200 3900 m 6 8 89 280 3600 142,000 30,000 2110.000 0 70 7b 75 68 75 70 70 m 2 148 60 85 91 111+ 139 1% * u 34 57 m 690 3200. 3700 4900 2 6 2h 12+ 450 7100 81,000 180,000 380,000 I"I'aken from plates snowing only h and 9 organisms. Log of cells per ml. Figure 2 Effect of various concentrations of ethyl violet on the growth curves of E, coli, in tryptose lactose broth*. 1-1OT dye l-50T dye 1-200r dye l-SOOT dye Time in Hours *Typical curves of at least 3 individual runs. 16 TABLE III Populations of E: coli in varying concentrations of ethyl violet in lauryl tryptose broth. Number of bacteria per ml. Concentration of dye 17 Hrs 1-101 1-50'1' ' " 1-lOOT 1-20021 1.1+00‘Tfi1u8‘031T Control 0 50 52 113 L11 56 56 50 H 2 59 78 90 103 83 103 101 E u 100 T110 4600 5900 4200 3600 2500 m 6 380 1510 «me 420,000 3140,000 220,000 121,000 a 1+6 39 37 1+3 1+3 #7 m 51 cu 2 37 92 88 102 105 102 123 E” 1+ 105 6900 5600 5100 6700 3800 3600 6 M30 u90,000 2h6,000 190,000 280,000 210,000 200,000 0 51 1+6 41; 37 1&1 51 1m to 2 1+7 97 83 108 97 102 109 E u 106 7300 #000 6300 6700 7700 2800 m 6 1430 610,000 320,000 370,000 250,000 360,000 260,000 Figure 3 Effect of various concentrations of ethyl violet on the growth curves of E, coli, in lauryl sulfate broth*. l-lOT dye l-SOT dye 1-100'1' dye l-E‘OOT dye l-SOOT dye Time in Hours I"Typical curves of at least 3 individual runs. 18 19 Lauryl tryptose broth produced some unexpected curves as seen in Table III and Figure 3, since the only major difference between this broth and the previous one is the presence of lauryl sulfate. In four out of the six dye concentrations studied, the dye—medium combination actually resulted in an increase in organisms per ml over the control. This stimulation was not observed in the most concentrated nor the most dilute solutions used. In the latter the curves coincided. At l-50T then, the dye-medium combination apparently offered a better medium, at least in this part of the growth curve, than the lauryl sulfate alone. At l-lOT the toxic properties of the dye are again strongly exerted, and although no attempt will be made at this time to explain this stimp ulation, it is of interest to note that the lauryl sulfate concentra- tion in this medium is also l-lOT, possibly indicating a molecular combination of some sort. Judging from the stimulation noted in the lag and log phases of the growth curves, lauryl tryptose broth was adopted as the base medium and 1-50T as the ethyl violet concentration for the comparison with brilliant green bile. However, in preparing the medium for use the dye was added before sterilizing. After the dye medium was sterilized and removed from the autoclave it was noted that 30 - 50 percent of the dye had pre- cipitated out of the medium. Autoclaving was attempted using several lesser dilutions of dye but the precipitation occurred in each case. Upon being added to the base medium and allowing it to stand for 24 hours, a small portion of the dye precipitated out but not to the extent encountered after autoclaving. Also, considering the fact that lauryl 20 sulfate might be used as a presumptive medium and therefore its results as a confirmatory medium might be open to doubt, the medium was removed from consideration. Turning back to the medium whiCh showed the next best results, tryptose broth, it was decided that a l-200T concentration of ethyl violet was too concentrated. The curve of the lag phase dropped slight- ly in the first two hours and the slope of the log phase was not quite steep enough to warrant its use, especially if weaker organisms might be encountered in field work. (Figure 2). It was thoughtthat l-NOOT was not concentrated enough, due to the similarities between it, the l-SOOT concentration and the control. This was later borne out by other work.(Figure u). A 1-3OOT ethyl violet concentration was finally chosen since it lay midway between that were considered the extremes. The medium was sterilized after the addition of the dye. After the autoclaved medium had stood for a few hmurs, a barely perceptible pre- cipitate was noticed. An experiment was then set up to determine the effect of precipitation of the dye on the growth curves of E; 3313, This test, incidentally, showed the toxicity curve from l-lOOT t0 1—400T and with interpolation, to l-SOOT. The test consisted of making up media, as in the previous growth curve determinations, and dividing the flasks into two groups. One group had the appropriate dye concentrations added while the Others were left alone; both groups were autoclaved simultanp eously. Dye was added to the second group of flasks just prior to use. An examination of Figure U indicates that autoclaving does have some ef- fect 0n the growth curve, and, while relatively slight, it is consistent. Log of Population at 6 hrs. 21 Figure 4 Effect of autoclaving ethyl violet tryptose broth medium.* (___) - I"Typical curves 6 l”. ,x)’ ” 5 h ---- Dye added post- autoclaving Dye added pre- autoclaving 3 H Eu 5+ 5+ 5+ 54 Ea o c> c> c3 c: a 2 .55. s E. s s I; I I I I I o .—I .-I H H H H O Dye concentration Line AB - Difference in population by adding dye before autoclaving (A) and after (B). Point C - Point at which population of medium adding dye before autoclaving is the same as point B in medium where dye is added after autoclaving. Line CD - Concentration of dye-medium which, when auto- claved, gives approximately the same results as l-3OOT dye added after autoclaving the medium. Portion of the curve extrapolated from Table II. of at least 3 individual runs. 22 Referring to Figure 4, points A and B on the graph show the differences in population when the dye was added prior to autoclaving and after autoclaving, respectively. Point C, therefore, represented the concen- tration of the dye added prior to autoclaving which gave the same results, in terms of population, as l-3OOT, added after autoclaving. A line drawn to the abscissa gives approximately 1-250T as the concentration yielding point C. This graph also shows the relative toxicities of the various concentrations. The ethyl violet became slowly more toxic from l-SOOT up to l-HOOT, edged downward slightly faster to ly3OOT and then suddenly became very rapidly toxic down to l-lOOT. It is assumed that this curve approaches the vertical at higher concentrations. These population figures were taken at six hours and the curve would probably be flatter at shorter time intervals (with the possible exception of N hours). The extreme end of the curve, from l-uOOT to l-EOOT Was inter- polated from Table II. It can be seen that l-3OOT lies near the top of the toxicity curve and may be a good choice. Tryptose lactose broth with a concentration of ethyl violet prior to autoclaving of l-250T, was finally chosen as the medium to be used in the field tests. Comparison with brilliant green lactose bile (BGB). Preliminary studies indicated that all presumptive tubes showing gas in 24 hours would confirm in the l-250T ethyl violet tryptose broth (EV) and were invariably coliforms, so it was decided to limit this portion of the study only to tubes inoculated from the same presumptive tube which did not agree on the presence of gas. The two parallel tubes of confirmatory 23 media were selected and attempts at recovering coliforms were made on both. "Standard Methods for the Examination of Water and Sewage" require that all presumptive tubes showing gas in 24 or #8 hours to be inoculated into a confirmatory medium, such as 3GB. Tubes of 3GB showing gas in 2H to #8 hours are then streaked on eosin methylene blue agar (EKB) plates and incubated for 24 hours. Typical and/or atypical coliform colonies are picked and inoculated into lactose broth and then streaked on an agar slant. Gram stains are made from the agar slant at 2H hours and if gas appears in the lactose tube in 2H or ”8 hours these prepara— tions are examined for the presence of spores or spore-forming bacilli. If the latter are present, they are inoculated into formats ricinoleate broth and if gas appears, the procedure is repeated beginning with EMB. In this particular study spore—formers were eliminated by the dilution plate method rather than using formats ricinoleate broth. This portion of the study was carried out on waters sent to the Michigan Department of Health from many parts of the state. As reflected in Table IV, over 80 percent of the samples were derived from wells, the rest coming from springs, city distribution systems, a lake and some from unknown sources. A total of luo samples was studied during the months of June, July, and August. Each sample was prepared by placing 10 ml into each of 5 tubes of double strength lactose broth and 1 ml into one tube of single strength lactose broth for a total of six tubes per sample. Of the total number of tubes in the samples 67.7 percent showed gas production, (Table V), or, were presumptive positive. 2% TABLE IV Source and number of samples. Source Number Wells ll7 Springs 2 Municipal distribution systems 10 Lakes 1 Unknown 10 Total 1&0 TABLE V The comparative confirmation of positive presumptive tubes by the use of brilliant green bile broth and ethyl violet tryptose broth =m := Medium. No. of tubes Percent Positive presumptives in lactose broth 569 07.7 Confirmed in Ben 47h 83.3 Confirmed in EV 486 85.4 Total no. of tubes in ‘ 1140 samples 840 25 A study of Table VI indicates that EV is slightly less toxic and allows the production of gas faster than the BGB medium, since 68.2 percent of the presumptive tubes confirmed in 2H hours, while 62.2 con— firmed in BGB in 2h hours. This represents an improvement of only six percent, but this increases slightly when broken down by times of pre— sumptive tubes showing gas. Here, 63.4 percent of the positive presump- tive tubes studied showed gas in NS hours and of these 51.1 percent confirmed in EV in 2h hours; an increase of 7.5 percent over BGB. Of the presumptive tubes showing gas in 2H hours the two confirmatory media were more nearly alike indicating that the EV medium was less toxic to the less hardy organisms encountered in the 24 - #8 hour period than was BGB, while the hardier organisms encountered in the first 24 hours were about similarly affected by both.media. The #8 hour group most likely contains the attenuated coliforms, coliforms whose lactose fer— menting power has been lost or otherwise affected and the slow lactose fermenting spore-forming bacilli. As seen in Table VIII the “8 hour group was of predominate importance in the differences of the two media. Before leaving this topic it might be noted that gas in the 2M hour EV medium appeared in larger quantities at this time indicating that it might also appear earlier. In the M8 hour group the amount of gas was approximately the same. Turning to the differences between EV and BGB in terms of samples and Most Probable Numbers (MPH), Table VII shows some interesting re- sults. Of the samples studied 33 showed differences of MPN and of these 21 gave EV a higher MPN than BGB while 12 gave lower MPN's. Further 26 mom :.mo Hem 9.0m mom mom :.mo Hon o.om mom masses 0.:H mm «.mm ow :.H m ~.oa mm :.:m mm n.m N esssaeeou poz m.~a mm 0.0m 3m m.H : H.Hm ems m.am mas m.m a .msn w: m.wo mam m.am awa ~.om How m.mo :mm m.:: and :.mm :ma .mun :m ammo ammo was w: psoo was :m pace peso .mun w: a“ pace .mus :m a“ nu tuom Hence 19mm .uon them .mom them Hence them o>apamom whom o>apwmom woesfimnoo .moum .monm abapmadmoam o>Hpmasmem cage 580nm >m o» consoanmne "Scum mam op compoundmpa .modmemm Had macaw opapfimom o>fipma§wmum mdfieooon we mean on dogmaou soapmauwweoo mo mafia H> mqmm m mm mm so He mam . >ms m ca aoH mmsw4 mom. one: no moamemm manna : momma m manna m opus A pass no .02 Hence an assume an seeeHn an smeeHn an smeuHa asom moHassm zms “asap moamamm Mo umpSsZ Mo umpadz mmamemm on» macaw honmmno>ww Azmnv nopadn oapmpoum umoa mo soapspanpman HH> wqm4e 28 Ha nonpfion ea .mnoo mops» * m.mm w.o: w.:w «.ma psooamm mm moH No m: mm om Nm 0H m mN mHspoa mm Nm mm Nm HN NH H HH em neon Nm mm aH mN 0.0m mo Nm mm mm on MN NH N OH mam .meoo moses mom HN :.NN ON mH a NN NH m z N N son and em eH omdoo mmpfiv * em so: NH H.NH :H HH m :H HH m o o 0 Es mam 5 .mnoo mops» * aH mndooo ammo .man w: .msa 1m Hmpoa .man w: .mnm 1m deuce .mun w: .man :m mean» no damn them H.309 £330 .330 .maoo .mnoo .maoo duos .580 we we * Hmpoe Hence .mmam .mg m: Bosh .mmam .mg :m Seam .m«mmmao .zmz aw mnwoouwm non moamamm amp macaw mQOfipmSustoo no cowudenpofio mafia HHHb mamde 29 breakdown shows that 28 of these samples differed by only one tube, 3 by two tubes and 2 by three tubes. Table XII adds a little more to this by indicating that these differences are fairly evenly distributed over the range of the number of positive tubes occurring in any one sample. In other words the occurrence of different MIT's was apparent— ly not exclusively dependent on the number of coliforms present in the water. Some importance might be attached to the fact that seven samples showed EV with 2 tubes positive and BGB with only 1 tube positive, however, the rest were farily evenly distributed. These 33 samples were analysed for time of confirmation. From the 24.nour positive presumptives no tubes confirmed in BGB that did not also confirm in EV, (Table VIII). The reverse, however, is not true, since a total of 4 tubes confirmed in EV which did not confirm in the corresponding BGB tube. A total of 12 tubes confirmed in both media from this group and a total of 11 tubes did not confirm at all. The importance of the 48-hour presumptive group is here emphasized; since 14 tubes confirmed in BGB and not in EV from this group while 22 con- firmed in Ev and not BGB and 53 confirmed in both media. This gives a total of lb tubes from the 24—hour positive presumptives and 89 from the 48-hour group. The major differences between the two media, then, are encountered with organisms that are slow lactose fermenters which may include attenuated coliforms and some aerobic and anaerobic spore- forming bacilli. able IX is a condensation of the above; and from the 33 nonpagreeing samples 40 tubes (in parallel) were studied with the intention of isolating coliforms from the media. 30 TABLE IX Distribution of tube differences among the non-agreeing samples w— ‘wfi “ Parallel tubes No; of tubes No. of Samples “ in Which; showing difference Represented in presence of gas *EV'+ and BGB - . 26 21 *EV - and 13GB + 14 12 Totals 40 33 *Note: only one classification occurred in any given samples 31 It is here noted that attempting to isolate coliforms from the presumptive medium by several different means would probably give a truer picture of the coliform population, however, this study has been confined to isolation according to the "Standard Methods" completed test and only from parallel tubes showing differences in gas production. Before discussing the tabulated results of this test it might be of interest to note a few observations made on the cultures isolated during the test. Of the 33 samples tested only one gave typical col- onies of E, ggli_and none of g, aerogenes on EMB. The remainder gave only atypical colonies. These were slow to develop being about one millimeter in diameter in 24 hours. After removing portions of colonies for the rest of the completed test, the plates were re-incubated.up to 72 hours. Two samples gave atypical colonies at 24 hours which slowly developed sheens to become typical E, ggli_colonies by 48 hours. Fur— ther study indicated that these were probably coliforms which had lost some of the power to ferment lactose and had become slow lactose fer- menters. Another sample yielded a large colony with a typical sheen at 48 hours (no sheen at 24 hrs.) and gave gas in lactose in 48 hours but when isolated in pure culture was a large gacillus probably of the B, aerosporus group. No further attempts at identification were made. IMViC reactions were determined on 6 to 8 atypical coliform cultures isolated from the samples and the predominate grouping seemed to be - + -‘+. however, no attempt was made to classify all of the cultures isolated. When the amount of gas in the inserts was compared it was found that EV produced up to two times as much gas, if from a 24 hour 32 presumptive, than BGB, although from the 48 hour presumptives the amounts were very similar. An examination of Tables X and XI shows the results of the attempt to isolate coliforms from the parallel tubes of confirmatory media showing different results. In this classification 26 tubes gave gas in EV and none in BGB while 14 tubes gave gas in BGB and none in EV. The completed test showed that coliforms were present in 14 out of the 26 parallel tubes, in EV as well as the negative BGB tubes. Two sets of parallel tubes gave inconclusive results, either due to errors in technique or some failure in the isolation. The rest failed to yield any coliforms. The first group can be considered BGB false negatives occurring as a result of the toxicity of the medium or failure to supply an adequate growth medium. The last group can be termed EV false positives and may be due to one or several factors, such as, 139 sufficient initial toxicity or after a large population of gram nega- tive organisms has removed sufficient dye to allow them to grow. The end result here, even though 10 EV false positives have oc- curred, is a net increase of 4 tubes of coliforms since 14 tubes havebeen uncovered by the Ev that would otherwise have been recorded as negative. The next category contains those tubes which gave no gas in Ev but did in BGB. Of the 14 sets of parallel tubes studied only 3 yielded coliforms in both tubes of the set. These can be considered EV false negatives. This is a big increase in recovery of organisms over the same category with BGB, and is a strong indication that the EV medium is less toxic to those weaker organisms. Ten sets of parallel tubes did not yield any coliforms and these can be considered.BGB false 33 TABLE x Results of completed test on samples showing difference in MPN from parallel tubes. No. of Tubes of EV‘+ and BGB - l"completed from: No. of Tubes (in parallel) (BGB false negative) Both B3B and EV 14 333 only 1 EV only 1 (EV false positive) neither 10 Total 26 No. of Tubes of Ev - and BGB + I"completed from: (EV false negative) Both BGB and EV 3 BGB only 1 EV only 0 (BGB false positive) neither 10 Total 14 Grand Total ‘ 40 *Hote; completed means coliforms isolated. 34 TABLE XI Occurrence of false tests as determined from Table X BGB EV Totals False +. 10 10 20 False - 1n 3 17 Totals 24 13 37 35 positives. This is the same number as that found with the similar group with EV. If the ethyl violet medium were less toxic and supplied a richer medium for the growth of attenuated coliforms, then it would be ex- pected to recover more of these organisms than BGB. Table XI shows this to be the case. A richer medium, on the other hand, would tend to allow the growth of interferers, especially the slow lactose fer. menting bacilli which form spores. Judging from Table XI it seems that, although the medium containing the ethyl violet dye was more nutritive, it allowed no more false positives among this group than the BGB. Furthermore it is suspected that the growth of these organisms causing the false tests in EV may be due to the removal of a portion of the dye by the much faster growing gram negative organisms normally present in water. The EV medium then, gave only 13 false tests to 24 for the BGB. This is a reduction of 54 percent for the sets of parallel tubes studied. Table XII presents the number of tubes showing gas in each sample and the medium which gave the correct results as determined by the comp pleted test. The ethyl violet was correct in 19 samples and 333 in 11. No pattern seems apparent as to the relationship of the number of tubes positive in a sample with the KPH, or whether EV gives a high or low MPN. TABLE XII Number of positive confirmatory tubes in each medium, by sample, and the medium showing the correct results by the completed test. No. of Samples No. tubes positive Medium correct by the completed test EV BGB EV BGB ? 4 6 a 3 1 l 6 l 1 5 4 1 2 4 3 2 l 3 2 l l 3 O l 7 2 1 2 4 1 2 2 O 2 2 l O 1 1 3 0 1 3 3 l 2 2 l l 2 3 l 1 2 4 1 2 3 4 1 1 1 4 5 1 l 5 6 l 19 11 3 KM KN SUKHARY AND COECLUSIOHS Ethyl violet dye at a concentration of 1-50,000 in lauryl tryptose broth showed the least interference with the lag and log phases of E, 9311, actually stimulating the organism, but could not be used due to excessive precipitation upon autoclaving or on standing. Studies with ethyl violet in tryptose broth.indicatedthat auto- claving has some effect on the activity of the dye and should be con- sidered if further studies are made. The medium decided upon for comparison with brilliant green bile is as follows: Tryptose 2.0% Lactose 0.5% KBHPOM 0.4% mercy, 0.153% NaCl 0. 53% Ethyl Violet l—250,000 final concentration added before sterilizing pH 6.8 before sterilizing Comparison with B9B indicated that the ethyl violet medium was less toxic to the weaker coliforms and equally toxic to the inter- ferers. It also allowed faster development of gas in the fermentation tubes allowing them to confirm earlier. It is therefore concluded that the 1-25OT ethyl violet tryptose broth shows great promise in the field of coliform confirmatory media 38 particularly due to its reduced toxicity to the weaker organisms. It is suggested however, that further studies be undertaken to overcome the false positive tests encountered. One method of attack might follow an observation made by Shane, that the number of false positives in B33 was greatly reduced by using lauryl sulfate broth as a presumptive medium instead of lactose broth. 10. ll. 12. LITERATURE CITED Boniece, J. R. and W. L. Hallmann. The optimum incubation temper- ature for the primary isolation of coliform organisms. Jour. Am. Water Works Assoc. 42:155, 1950. Bordt, D. E. A study of chemical agents for selective growth of the coliform organisms. Unpublished master's thesis. East Lansing: Hichigan State College. 1951. Churchman, J. W. The selective bactericidal action of gentian violet. Jour. of Exptl. Med. 16:221. 1912. Darby, C. W., and W. L. Hallmann. Studies on media for coliform organisms. Jour. Am. Water Works Assoc. 31:689. 1939. Darby, C. W. Studies on primary and selective media for colifonm organisms. Unpublished Haster's thesis. East Lansing: Michigan State College. 1943. France, R. L. A comparative study of certain presumptive media for testing raw water. Jour. Am. Water Works Assoc. 28:785. 1936. Hale, F. E. Discussion of brilliant green bile media. Stain Tech. 2:2N. I927. Jolliffe, E. M. Studies on the effect of ethyl purple on certain bacteria. Unpublished Haster's thesis. East Lansing: Hich- igan State College. 1998. Jordan, H. E. Brilliant green bile for the determination of the colon-aerogenes group. Jour. Am. Water Works Assoc. 18:337. 1927. Litsxy, W., W. L. Mallmann, and C. W. Fifield. Ethyl violet. A selective dye for the isolation of gram negative bacteria. Stain Tech. 27:229. 1952. Hallmann, W. L. and J. M. Hepler. A comparative study of Standard Methods of Water Analysis (1933) and two percent bile brilliant green lactose broth confirmation. Jour. Am. Water Works Assoc. 28:“11. 1936. Mallmann, W. L. and C. W. Darby. Uses of lauryl sulphate lactose broth for the detection of coliform organisms. Am. Jour. of Publ. Health. 31:127. 1941. fill. llllll I‘ll" 1| Ilalll“ . _ n a v p c I u o I . . . . . _ u . e e . . . . . r . . o u o . o A . . . e o n . . . . . o . ¢ - a s u r . c a . A . . . 4 . . A . . u - . . n I . s e u . - . v 5 o . a u e v n p v a 9 v . . . . . . o I I u . - I u o u . e . I u I e 0 . u . a b u t 7 . b 13. in. 15. lb. 17. 18. 19. 20. 21. 22. 25. 40 HcCrady, M. H. A comparison of hacConkey's broth and standard lactose broth as media for detection of coliform organisms in water. Am. Jour. of Publ. Health 29:1250. 1939. Huer, T. C. and R. L. Harris. value of brilliant green in elim— inating errors due to the anaerobes in the presumptive test for E, coli. Am. Jour. of Publ. Health 10:874. 1920. Petroff, S. A. and W. S. Sump. Bacteriostatic and bactericidal studies of various dyes and allied compounds. The Jour. of Lab. and Clin. Med. 20:689. 1935. Prescott, S. C., C. E. A. Winslow, M. H. McCrady. "Water Bacter- iology", 6th edition, 1996. John Wiley and Sons, New York. Richey, D. Relative value of 2% and 5e brilliant green bile cons firmatory media. Jour. Am. Water Works Assoc. 33:649. 1941. Ruchhoft, C. C. Comparative studies of Standard Kethods and the brilliant green bile medium on Lake Michigan Water at Chicago. Jour. Am. Water Works Assoc. 16:778. 1926. Ruchhoft, C. C. and J. F. Norton. Study of selective media for 0011- Aerogenes isolation. Jour. Am. Water Works Assoc. 27:113“. 1935- Salle, A. J. The differentiation of the colon—aerogenes group of bacteria. Jour. Am. Water Works Assoc. 21:71. 1929. Salter, R. C. Observations on the rate of growth of E, coli. Jour. Inf. Dis. 24:260. 1919. Shane, M. S. Studies on false confirmed tests using brilliant green bile broth and comparison studies on lauryl sulfate tryptose broth as a presumptive medium. Jour. Am. Water Works ASSOC. 39:337. 1947. Standard Xethods for the Examination of Water and Sewage, 9th edition. 1946. Am. Publ. Health Assoc. New York. Wattie, E. Coliform confirmation from raw and chlorinated waters with brilliant green lactose broth. Publ. Health Reports 58:377. 19h}. Winslow, C—E. A. and A. E. Doloff. The relative effect of certain triphenyl methane dyes upon the growth of bacilli of the colon group in lactose broth and lactose bile. Jour. of Inf. Dis. 31:302. 1922.