Isolation of Salmonellae by Selective Motility Systems

Transcription

1 APPLIED MICROBIOLOGY, May, 1965 Copyright 1965 American Society for Microbiology Vol. 13, No. 3 Printed in U.S.A. Isolation of Salmonellae by Selective Motility Systems PHILIP F. STUART AND HILLIARD PIVNICK Connaught Medical Research Laboratories, University of Toronto, Toronto, Ontario, Canada Received for publication 16 December 1964 ABSTRACT STUART, PHILIP F. (University of Toronto, Toronto, Ontario, Canada), AND HILLIARD PIVNICK. Isolation of salmonellae by selective motility systems. Appl. Microbiol. 13: Salmonellae were isolated from naturally infected swine and human feces by means of selective migration through semisolid enrichment (SM) media in modified "U" tubes. Comparative studies showed that recovery of Salmonella by SM techniques was equal or superior to that of standard procedures employed in two routine diagnostic laboratories. Primary Salmonella isolations from SM enrichment were relatively free from normal fecal bacteria and often pure. The SM method required only 15% of the media and correspondingly less time and skill than the usual methods for isolating Salmonella. Two basic procedures for the isolation of salmonellae from feces have gained wide acceptance. In one, a portion of the specimen is plated directly onto one or more solid selective media. In the other, the specimen is enriched in a liquid medium which favors preferential growth of salmonellae. At intervals, the enrichment culture is plated on solid selective media. Normal fecal bacteria which grow on the plates in both procedures may overgrow, mimic, or inhibit salmonellae. For this reason, several colonies suspected of being Salmonella must be carefully picked from each plate and grown in tubes of differential media for identification. Experience has shown that successful isolation of salmonellae is best accomplished by using a combination of several plating and enrichment media for the examination of each specimen (Edwards and Ewing, 1957), with the hope that Salmonella colonies will be clearly defined in at least one. This complex examination requires many units (tubes and plates) of enrichment and differential media and much skilled technical manipulation. Numerous attempts have been made to devise simpler, more efficient methods for isolating salmonellae. Many of these attempted to improve Mueller's (1923) tetrathionate or Leifson's (1936) selenite enrichment media. Rappaport, Konforti, and Navon (1956) described a new enrichment medium which was superior for nontyphoid salmonellae. Cultures of infected fecal specimens enriched in Rappaport's medium frequently grew pure salmonellae when plated on agar. To achieve the high degree of purity, however, the fecal specimen had to first be diluted 1:1,000 before inoculation, thus reducing the chance for detecting salmonellae present in small numbers. The other disadvantage of Rappaport's medium resulted from its selective agent, malachite green, which was completely inhibitory to Salmonella typhosa and even partially inhibitory to certain other Salmonella species. Confirmation of the advantages and disadvantages of Rappaport's medium can be found in papers by Rappaport, Skariton-Loewenthal, and Olitzki (1953); Rappaport et al. (1956); Rappaport and Konforti (1959); Collard and Unwin (1958); and Iveson, Kovacs, and Laurie (1964). Several methods have been devised for isolating salmonellae on the basis of their motility 365 (W.L. Mallmann, personal communication), and two recent references have been concerned with motility through agar. Jones and Handley (1945), while reporting a new liquid enrichment medium, described the incidental application of a selective motility medium with which they successfully isolated salmonellae from necropsy specimens overrun with intestinal bacteria. Ino and Graber (1955) reported a method for separating Salmonella cultures contaminated with Pseudomonas by inoculating a "U" tube containing a nonselective semisolid medium with the bacterial mixture. The less aerophilic Salmonella penetrated the medium and completed migration sooner than did the Pseudomonas. We felt that enrichment media combining selective growth and preferential motility of salmonellae would facilitate their recovery from mixtures with competitive organisms. The media we devised were called selective motility (SM) media.

2 366 STUART AND PIVNICK A PPL. MICROIBIOL. RECOVER 7 mm. TUSI DIAMETER RV SITE FIG. 1. LOOSE CAPS FL w If I. Selective motility tube. < F INOCULUM SITE SM MEDIUM MATERIALS AND METHODS Preliminary trials with SM media were done in the motility tubes of Craigie (1931). For several reasons, these were unsatisfactory and were replaced with modified "U" tubes (SM tubes) in which a small-bore side arm tube was attached to the base of a test tube (diameter, 2.54 cm), as shown in Fig. 1. The SM tubes permitted inoculation of a large test specimen onto the semisolid medium in the main bore. Motile bacteria grew to the bottom of the main bore and could be readily seen migrating to the surface of the medium in the side arm. Cultures of migrant organisms from the side arm were plated onto solid media and examined. The relative percentage of Salmonella colonies and the per cent and identity of competitive species were recorded. A large number of different selective semisolid (SM) media were screened for effectiveness in recovering five common Salmonella serotypes from artificially infected feces. The effect of several physical variables such as incubation temperature, depth of the medium, and ph adjustment were also tested. Promising media were retested for the ability to isolate salmonellae from specimens inoculated with as few as 3 to 10 pathogens per gram of feces. Media and methods which proved satisfactory in these preliminary experiments were tested for practical efficiency in isolating salmonellae from naturally infected human or swine fecal specimens. Parallel examinations of the same samples were done by independent workers employing the standard procedures outlined in the introduction. Three different SM media were chosen for extended study, and their composition will be described. Selective motility selenite (SMS) medium. A volume of double-strength Selenite F Broth (BBL) heated to boiling was mixed with an equal volume of autoclaved 1.2% agar in water. Brilliant green dye (Difco), 1:1,000 aqueous, was added to a final concentration of %. The medium was dispensed hot to a depth of 5 cm in the SM tubes and allowed to solidify at 4 C. SMS medium was made fresh for most test series but could be held up to 3 days in a refrigerator before use. Selective motility Rappaport (SMR) medium. Agar (0.6%) was dissolved in the enrichment medium of Rappaport et al. (1956), which is composed of Tryptone (Difco, 0.5%), NaCl (0.8%), KH2PO4 (0.16%), malachite green (0.012%; British Drug House, Toronto, Canada), and MgCl2.6H20 (4%). The medium was dispensed to a depth of 5 cm in SMR tubes and autoclaved. SMR medium was indefinitely stable at 4 C. Brilliant green modified Rappaport (SMBGR) medium. SMBGR medium was made, dispensed, and sterilized exactly like SMR medium, except that the malachite green was replaced by brilliant green (0.001%). SMBGR medium was also stable when refrigerated. Preparation of specimens. Concentrated (two to four times) solutions of each of the above media without agar were prepared and stored within the limits of their stabilities. Fecal specimens in transport fluid were mixed with a sufficient quantity of concentrated liquid media to render a fluid inoculum in which the final medium concentration was approximately single strength. About 2 ml of this inoculum was poured onto the semisolid medium in the main bore of the SM tube. In one test series (Aberdeen), in which the fecal specimens were received in the natural state, 2 ml of single-strength liquid medium were dispensed into the main bore of the SM tube, and the fecal inoculum was added with an applicator stick. The inoculated tubes were incubated at 37 C until there was visible growth at the top of the medium in the side arm. At this time, a loopful of the culture in the side arm was streaked onto EMB agar (Difco). The EMB plates were incubated overnight at 37 C and examined. Plates with colonies resembling salmonellae were retained, and the rest were discarded. One typical colony from each of the retained plates was cultured overnight in a Triple Sugar Iron (TSI) Agar (Difco) slant, and the plates were stored in a refrigerator. TSI-agar cultures which appeared to be salmonellae were further tested biochemically and serologically (Edwards and Ewing, 1957). If the TSI agar culture proved not to be Salmonella, the stored EMB plates were reviewed and, if necessary, additional colonies were picked. Colonies on EMB agar which resembled salmonellae on the first inspection, but were not Salmonella, often acquired a distinctive appearance during storage which obviated retesting. In a few cases, however, it was observed that, at the first appraisal, a colony which was not Salmonella had been cultured in TSI agar from a plate on

3 VOL. 13, 1965 TABLE 1. Efficiency of isolating salmonellae from 246 swine fecal specimens (series 1-SMS) Method No. of Per cent of isolations all isolations Standard (12)* 61.0 SMS (28) 83.0 SMS + standard * Numbers in parentheses indicate the number of Salmonella isolations by one method only. which salmonellae were present, and that refrigerator storage had made proper identification obvious. SM tubes in which no bacterial migration had occurred by the seventh day of incubation were considered negative for the presence of motile salmonellae and discarded. ISOLATION OF SALMONELLAE RESULTS All the following data compare the SM techniques with the standard methods for isolation of salmonellae. Comparative data concern the efficiency of recovery, purity of isolates, time requirements, and economy of material and labor. Efficiency of the SM technique refers to its ability to detect salmonellae in infected fecal specimens. Data on SM efficiency (Tables 1, 5, and 6) are grouped as follows: (i) detection of salmonellae from specimens known to be infected and (ii) detection of salmonellae from routine specimens which may or may not be infected. The expression "purity of the SM isolates" refers to tne percentage of Salmonella colonies present on each EMB Agar plate streaked from 367i SM enrichment of an infected specimen. Plates which contained only Salmonella colonies were considered to be "100%" or "pure" cultures; plates which were not "pure" were rated according to the percentage of Salmonella colonies present on each (99 to 50, 49 to 10, and < 10% Salmonella colonies, respectively). Table 2 shows the number and per cent distribution of the Salmonella-positive plates in each series grouped according to the ratings above. Species of normal fecal bacteria which survived and migrated in SM enrichment varied with the type of SM medium. Usually, a single species would predominate in SM enrichments which were not infected with salmonellae, although, occasionally, up to three distinct types could be found on the EMB Agar plate. Salmonella isolates which were not pure usually had only one other species associated. The organisms, not Salmonella, which completed SM migration were, in order of prevalence, Proteus, Escherichia, Aerobacter, and Paracolobactrum in SMS medium, and Escherichia, Aerobacter, and Paracolobactrum (but no Proteus) in SMR and SMBGR media. On EMB Agar these species were readily differentiated from salmonellae. Table 3 shows the time required for salmonellae to migrate from the site of inoculation to the recovery site in the side arm. Table 4 compares the number of units (plates and tubes) of each type of medium required for the examination of 100 routine human fecal specimens by the standard and the various SM procedures. Series 1. SMS medium with swine feces. A series of 246 fecal specimens from a herd of swine naturally infected with at least three different serotypes of salmonellae was examined in SMS medium. The specimens were received refrigerated in the natural state in lots of 10 to 20 over a 6-month period. Parallel examination of the specimens was done by an independent group who plated both the fresh feces and Tetrathionate Broth (Difco) enrichment cultures onto Mac- Conkey Agar and Brilliant Green Agar (both Difco). The efficiency of the SMS and standard systems in isolating a total of 72 salmonellae from the 246 specimens is shown in Table 1. SMS detected 28 salmonellae not recovered by the standard procedure and, conversely, missed only 12. Of the 60 salmonellae isolated from swine feces by SMS enrichment, 68% were in pure culture (Table 2, series 1), and the remainder, with two exceptions, were present in sufficient numbers on the EMB Agar plates to make selection easy. For this series, we used 7 cm of medium in the SM tube. Subsequently, we found that 5 cm was optimal. Probably because of the increased distance of migration, SMS series 1 had a higher percentage of pure isolates (Table 2) and a longer migration time (Table 3) than SMS series 2. Series 2. SMS medium with human feces. A preliminary series of 115 selected fecal specimens preserved in glycerol-saline were obtained from a public health laboratory which had examined them several days to several weeks previously. Using standard methods, they found that 57 specimens contained salmonellae and 58 (lid not. After examining the samples with SMS medium, we learned from the public health laboratory that we had detected salmonellae in 41 of their 57 positive specimens and in 1 of the 58 presumed negative samples for an efficiency of 73%. Over a 2-month period, 1,181 unselected routine specimens preserved in glycerol-saline were acquired from the same laboratory within 2 days after standard examination, and we re-examined them using the SMS enrichment method for com-

4 368 STUART AND PIVNICK APPL. MICROBIOL. Salm o nella colonies on EMB plate TABLE 2. Purity of Salmonella isolates plated from SM system Series I-SMS plates Series 2-SMS plates Series 3-SMR plates Series 4-SMBGR* plates l No. Per cent No. Per cent No. Per cent No. Per cent 100 (pure) < Total * Including S. typhosa isolates. parison. Efficiency of the standard and SMS procedures in isolating a combined total of 126 salmonellae was equal (Table 5, series 2). SMS enrichment detected 14 salmonellae missed by the standard procedure and, conversely, missed 15. The SM system is more economical than standard methods. Standard testing at the public health laboratory included direct plating of the specimens onto Wilson and Blair, MacConkey, and desoxycholate-citrate agars, and enrichment in Selenite F followed by plating on Wilson and Blair and Salmonella-Shigella agars. From the various plates, an average of 10 colonies per specimen were cultured in TSI Agar slants to identify them as possible salmonellae or shigellae. Thus, a total of 18,896 units of medium was required for standard examination of the 1,181 samples. In contrast, SMS testing required 1,181 tubes of SM medium, 1,075 EMB Agar plates, and 371 tubes of TSI agar-a total of 2,627 units of medium. This was less than 15% of the units of medium necessary in the standard examination (Table 4). It must be noted, however, that Shigella testing was included in the standard examination. Of the salmonellae isolated from human feces by SMS enrichment, 84.2% were pure enough to be easily seen (Table 2, series 2). Fewer than 5% of the isolates in the series were difficult to detect on EMB Agar, usually because they were overgrown by aerobacters. Series S. SMR medium. SMR enrichment was used to examine 424 routine human fecal specimens which were tested concurrently by the public health laboratory with the standard methods outlined in series 2. Parallel testing with SMS and Rappaport liquid enrichments was performed with 271 and 304 of these specimens, respectively. SMR medium was highly efficient in recovering salmonellae of 14 serotypes (Table 5, series 3), and SMR was superior in efficiency to both SMS and Rappaport's liquid in limited trials (Table 6). SMR enrichment yielded a higher percentage TABLE 3. Time required for salmonellae in naturally infected feces to migrate through SM media Series Medium No. of Days for migration salmo -_ Type Depth isolated Range Avg cm 1 SMS SMS SMR SMBGR 5 29* SMBGR 5 St * Nontyphoid Salmonella. t S. typhosa. TABLE 4. Average units of medium used per 100 specimens examined Method Enrich- Plating De ment Paig eniali Total Standard* ,000 1,600 Rappaport's liquid SMS SMR SMBGR * Includes examinations for Shigella. of pure cultures of salmonellae than did any other system (Table 2, series 3). Migration time in SMR (Table 3, series 3) was the greatest for any of the SM media dispensed to a depth of 5 cm, and medium requirements (Table 4) were the least. Series 4. SMBGR medium. SMBGR medium was assessed in several short trials. It proved superior in efficiency to standard procedures in isolating salmonellae from a limited number of swine fecal specimens. In the examination of 18 human fecal specimens which were known to be infected, SMBGR medium was 100% efficient.

5 VOL.. 13, 1965 ISOLATION OF SALMONELLAE 369 TABLE 5. Comparison of efficiency of SM and standard procedures in isolating Salmonella 8erotypes from human feces No. of isolates "O"01Y anti- Serotype Series 2 Series 3 Series 4 gen group ype (1,181 specimens) (424 specimens) (388 specimens) Standard SMS Standard SMR Standard SMBGR B S. heidelberg 6 6 S. paratyphi B S. paratyphi B var. java 6 (1)* 5 S. st. paul 7 (3) S. typhimurium 35 (6) 34 (5) 13 (2) Untyped 0 3 (3) 0 2 (2) Cl SS. choleraesuis 0 1 (1) S. infantis 2 4 (2) 2 3 (1) 2 2 S. montevideo (1) 1 S. manhattan 1 1 S. oranienburg 2 2 S. richmond S. thompson (1) Untyped 0 3 (3) 0 2 (2) 0 1 (1) C2 S. blockley S. muenchen 3 (1) 2 1 (1) 0 S. newport 10 (1) Untyped 0 1 (1) D S. enteritidis 3 3 S. dublin 1 3 (2) Untyped 1 1 E S. senftenberg 2 2 G S. poona 2 (2) 0 Untypable 1 (1) 1 (1) Total recovered by each method 112 (15) 111 (14) 32 (5) 34 (7) 14 (0) 17 (3) Efficiency of methodt * Numbers in parentheses indicate the number of Salmonella isolations by one method only. t Efficiency of method was determined by the following formula: total recovered by each method X 100 total recovered by all methods TABLE 6. Comparative efficiency of standard, Rappaport's liquid, and SM enrichments Salmonellae isolated by each method lrpqrid SMS SMR SMBGR -* 8 13 (4)t (1) (1) * Not done. t Numbers in parentheses indicate the number of Salmonella isolations by one method only.

6 370 30STUART AND PIVNICK APPL. MICROBIOL. TABLE 7. Efficiency of Salmonella typhosa isolations from 235 specimens by SMBGR and standard methods No. of Per cent Method isolations of total isolations Standard.28 (22)* 93.0 SMBGR.8 (2) 27.0 Standard + SMBGR * Numbers in parentheses indicate the number of S. typhosa isolations by one method only. Testing of SMBGR with routine human fecal specimens for comparison of efficiency with standard methods was done in two trials. One test was performed on 153 specimens received from the same source as those in series 2 and 3. The other test was conducted on 235 specimens in Aberdeen, Scotland, during the typhoid epidemic of May through June 1964, primarily to test SMBGR efficiency in isolating typhoid bacilli. Standard examination in the Aberdeen trial consisted of direct plating of whole specimens onto desoxycholate-citrate agar and enrichment of the specimens in Selenite F broth followed by plating onto MacConkey agar (all Oxoid). SMBGR enrichment was superior to the standard methods in both of these trials (Table 5, series 4); a total of 17 nontyphoid salmonellae were isolated for an efficiency of 100%. In testing the series of 153 specimens, SMBGR was more efficient than all of the other techniques combined (Table 6). Isolation of S. typhosa in the Aberdeen series was only 27 % efficient (Table 7). The typhoid bacilli isolated, however, were all of a single phage type, no. 34, so that a true picture of efficiency in a mixed typhoid population could not be assessed. Purity of the SMBGR isolates both for S. typhosa and nontyphoid salmonellae was second only to isolates from SMR enrichment (Table 2, series 4). Nontyphoid salmonellae migrated faster in SMBGR than they did in any of the other SM media (Table 3). Even the maximal time required for isolation of salmonellae in this medium fell within the normal limits for standard enrichment S. typhosa migrations were, however, somewhat slower, and SMBGR tubes had to be incubated the full 7 days required by other SM media for best results. Although SMBGR enrichment proved slightly less economical to use than did SMR or Rappaport's liquid (Table 4), the advantage of superior efficiency and speed far outweighed this fact, and SMBGR was, like all SM media, much less expensive to employ than the standard. DIscussIoN SM systems are effective because, unlike convential liquid enrichment media, the semisolid medium provides a continuously fresh enrichment environment for motile bacteria. As in other enrichment media, preferential multiplication of salmonellae occurs in the presence of selective nutrients and inhibitors. However, in the SM system, active migration in the medium eliminates nonmotile species from the start and provides any motile salmonellae present with a constantly renewed growth advantage over other motile bacteria. By the time migration has reached the side arm of the SM tube, the concentration of salmonellae is usually greater than that of other organisms, and often salmonellae are the sole survivors. Since bacterial migration is visible, the optimal time for plating of the enrichment culture is not subject to question as is the case in liquid media, and EMB Agar plates streaked from SM enrichment often grow pure or nearly pure cultures of salmonellae when any are present in the fecal inoculum. Because of the purity of the salmonellae isolated and the efficiency of the systems, SM procedures require less medium, less technical skill, and fewer work hours than do standard procedures. A single technician with no previous experience with salmonellae did all of the work in these trials. Despite this economy, the SM methods were at least as effective as standard methods. Nonmotile salmonellae are rare, but they do occur and are unsuitable for SM enrichment, as are shigellae. Typhoid bacilli are less motile than other salmonellae, and, though they were effectively isolated with 27% efficiency in SMBGR medium and occasionally isolated in SMS medium, the SM examination was not the system of choice for typhoid diagnosis. We suggest that the SM examination be accompanied by a single direct plating of fecal specimens onto a selective medium such as desoxycholate or Salmonella- Shigella agar. The additional step would detract little from the economy of effort and material achieved by the SM procedure and would allow a good chance for detection of nonmotile salmonellae, increase efficiency of S. typhosa isolations, and provide the method of choice for detection of shigellae. Although the theoretical mechanism of SM enrichment was the same in each case, the performances of the three SM media varied somewhat. Each will be discussed below.

7 VOL. 13, 1965 ISOLATION OF SALMONELLAE 371 The results of the trials with SMS medium were presented mainly to demonstrate the general value of the SM system in a large testing series. Even this medium, which gave the poorest results of the three SM media, has been shown to be as efficient as the standard methods in detecting salmonellae and greatly superior with respect to ease and economy of operation. The obvious imperfections of SMS medium are its instability to storage and autoclaving and the lengthy time of migration of salmonellae through it. Salmonellae isolated from SMS enrichment, though much freer from organisms that are not Salmonella than isolates from most liquid enrichment, were less often pure than those from the other two SM media. SMR medium embodies the extreme selectivity of Rappaport's liquid enrichment and is useful for the detection of nontyphoid salmonellae. The efficiency of Salmonella isolation in SMR medium was excellent, and the isolates were usually in pure culture. SMR was so selective that only 60% of the SM tubes inoculated showed growth in the side arm and required streaking. For this reason, economy was greater than for any of the compared systems. SMR medium proved to be more efficient and less complicated to apply than its liquid analogue. Elimination of the 1:1,000 dilution of the fecal inoculum required for successful application of Rappaport's liquid enrichment reduced labor and probably allowed detection of lower concentrations of salmonellae. SMR enrichment has, however, some disadvantages. Because of partial inhibition by malachite green dye, salmonellae required a longer migration time than they did in other SM media, and the dye is totally inhibitory to S. typhosa. This made SMR less suitable than other media for the examination of human excreta. We suggest that SMR medium might prove valuable in the diagnosis of veterinary salmonellosis where speed is of less importance, typhoid is nonexistent, and efficiency, economy, and ease of manipulation would be prime considerations. Avian examinations in which S. pullorum is a factor would be obvious exceptions. In 1956, Rappaport et al. (1956) reported that replacement of the ultraselective malachite green in their liquid enrichment with any other inhibitory agent resulted in decreased efficiency. We found that brilliant green dye could be effectively substituted for malachite green in SMR medium. This was probably possible because the inherent selectivity of the SM system augmented the residual selectivity of the Rappaport medium modified with brilliant green. The resulting medium, SMBGR, was superior in many ways to SMR medium. The reduced toxicity of the dye in SMBGR appeared to allow freer growth of all types of salmonellae while preserving most of the inhibitory action of SMR medium toward species that were not Salmonella. As a result, SMBGR medium proved more efficient than any other system in isolating nontyphoid salmonellae, and it showed a limited ability to recover S. typhosa. Migration times for salmonellae in SMBGR were much shorter than those in SMR or SMS. This, too, was a result of the low toxicity of the medium. SMBGR medium allowed more species that were not Salmonella to migrate than did SMR. For this reason, salmonellae isolated from SMB- GR were in pure culture less often than those from SMR, and more plating and differential media were required to process the extra migrants. SMBGR was, however, superior to SMS in both of these respects. During the Aberdeen trial, it was found that S. typhosa recovery from SMBGR was enhanced by deleting the phosphate buffer and raising the ph of the medium from 5.5 to 6.9. In 1953, Rappaport et al. (1953) stated that the acid condition was essential to suppress Proteus. We noticed no increase of Proteus migrations in the unbuffered medium. SMBGR medium has been effectively applied in limited trials for the detection of Salmonella in animal feeds and sewage. We did not, however, ascertain whether more than one serotype of Salmonella was present in any of these specimens or what the behavior of several strains in a single SM tube would be. SM techniques offer many advantages in Salmonella diagnostic work. Of the three SM media discussed, SMBGR appears to be the most promising for general use. Although we found SMBGR medium to be 100% efficient for paratyphoid species in our limited trials, more extensive testing must be done, and the problem of S. typhosa isolations will require more work. ACKNOWLEDGMENTS We wish to thank the following people: S. A. Horvath and C. Hughes for excellent technical assistance; M. Finlayson of the Ontario Department of Health and A. Rokitnicki of our laboratory for specimens, standard testing data, and cooperation; R. Cruickshank and A. MacDonald for courtesy and assistance extended during the work in Aberdeen; and W. L. Mallmann, J. H. Silliker, and F. S. Thatcher for criticism of the manuscript. LITERATURE CITED COLLARD, P., AND M. UNWIN A trial of Rappaport's medium. J. Clin. Pathol. 11:

8 372 STUART ADN] D PIVNICK APPL. MICROBIOL. CRAIGIE, J Studies on serological reactions of flagella of B. typhosus. J. Immunol. 21: EDWARDS, P. R., AND W. H. EWING Identification of Enterobacteraciae. Burgess Publishing Co., Minneapolis. INO, J., AND C. D. GRABER Recovery of Salmonella from contaminated cultures. U.S. Armed Forces Med. J. 6:586. IVESON, J. B., N. KovAcs, AND W. LAURIE An improved method of isolating salmonellae from contaminated desiccated coconut. J. Clin. Pathol. 17: JONES, R. E., AND W. R. C. HANDLEY A selective medium for the isolation of Salmonella from heavily contaminated material. Monthly Bull. Min. Health (Gr. Brit.) 4: LEIFSON, E New selenite enrichment media for isolation of typhoid and paratyphoid (Salmonella) bacilli. Am. J. Hyg. 24: MUELLER, L Un nouveau milieu d'enrichissement pour la recherche du Bacille typhique et des paratyphiques. Compt. Rend. Soc. Biol. 89: RAPPAPORT, F., AND N. KONFORTI Selective enrichment medium for paratyphoid bacteria. Appl. Microbiol. 7: RAPPAPORT, F., N. KONFORTI, AND B. NAVON A new enrichment medium for certain salmonellae. J. Clin. Pathol. 9: RAPPAPORT, F., M. SKARITON-LOWENTHAL, AND A. L. OLITZKI A culture medium for the enrichment of salmonellae and growth inhibition of Proteus. Bull. Res. Council Israel 2: Downloaded from on December 5, 2018 by guest