APPLIED AND ENVIRONMENTAL MICROBIOLOGY, OCt. 1983, p. 840-845 Vol. 46, No. 40099-2240/83/100840-46$02.00/0Copyright © 1983, American Society for Microbiology

GELRITE as an Agar Substitute in Bacteriological MediaDANIEL SHUNGU,* MARY VALIANT, VICKI TUTLANE, ELLEN WEINBERG, BARBARA

WEISSBERGER, LARRY KOUPAL, HANS GADEBUSCH, AND EDWARD STAPLEY

Clinical Microbiology Services, Merck Institute for Therapeutic Research, Rahway, New Jersey 07065

Received 23 June 1983/Accepted 26 July 1983

GELRITE gellan gum (formerly known as PS-60 and S-60) is a new naturallyderived, highly purified polysaccharide which displays several interesting proper-ties, including selfgelling. The suitability of GELRITE as an agar substitute wastested by evaluating the performance of several media selected from among thosemost commonly used in the isolation, identification, and enumeration of microor-ganisms in clinical laboratories. Fifty different bacterial species previouslyimplicated in human infections served as test strains. On the basis of the variousparameters considered, namely, colony characteristics, biochemical reactions,hemolytic patterns, and plating efficiency, media gelled by agar and by GELRITEcompared quite favorably.

GELRITE gellan gum (formerly known as PS-60 and S-60) is a new naturally derived (afermentation product of a Pseudomonas spe-cies), highly purified heterosaccharide biopoly-mer discovered by Kelco Div., Merck & Co.,Inc., San Diego, Calif. Its most prominent char-acteristics include ability to gel in the presenceof a cation, agar-like rigidity, compatibility withnutrient additives, thermal stability, and opticalclarity (1, 2). The present study was undertakenwith a twofold objective in mind: to determinethe effect of GELRITE in comparison with agaron bacterial growth as judged by plating efficien-cy and to evaluate the performance of this gellangum as an agar substitute in several mediacommonly used for routine isolation and identifi-cation of bacterial species from clinical speci-mens.

MATERIALS AND METHODSBacterial species. Fifty different bacterial species

representing a total of 26 genera were tested (seeTables 3 and 4). These cultures were chosen from acollection of approximately 2,600 clinical isolates ofhuman origin acquired within the past 2 years fromhospitals across the United States and kept frozen at-70°C. Before use, the test strains were regrown on ablood-based medium to check for purity.Media. Culture media and ingredients used in this

study are identified in Table 1. Freshly prepared platesof tryptic soy agar and tryptic soy GELRITE contain-ing 5% sheep blood were provided by GIBCO Diag-nostics, Madison, Wis. The following dehydrated me-dia without the gelling agents were custom-made forKelco by the indicated manufacturers: bismuth sulfite(BBL Microbiology Systems, Cockeysville, Md.), bril-liant green (BBL), eosin methylene blue (Acumedia,Baltimore, Md.), MacConkey (BBL), salmonella-shi-gella (Acumedia), and triple sugar iron (Acumedia).

All the other media and ingredients were taken fromour own laboratory stocks as purchased from regularsuppliers.Media containing agar or GELRITE as the solidify-

ing agents were freshly prepared in accordance withboth published (Kelco Commercial Development bul-letins CD-26 and CD-27) and unpublished informationprovided by Kelco (Table 1). An important consider-ation in the formulation of these guidelines was theachievement of comparable gel strengths (grams persquare centimeter) between agar- and GELRITE-con-taining media. The basic procedure for the preparationof the various media entailed the following: (i) dissolu-tion of specific quantities of dehydrated medium,MgCl2 * 6H20, and GELRITE or agar in distilledwater; (ii) autoclaving the dissolved mixture of ingredi-ents for 15 min at 121°C and 15 lb/in2 in a small bench-type agar sterilizer (New Brunswick Scientific Co.,Edison, N.J.); (iii) cooling the medium down to aholding temperature of 60°C (GELRITE) or 50°C(agar) after a gentle swirl; and (iv) dispensing themedium into either standard 15-mm petri dishes at 20ml per plate or screw-capped tubes at 7 ml per tube(for triple sugar iron). The preparation of both salmo-nella-shigella and bismuth sulfite media did not includeautoclaving. In each case, the medium mixtures weresimply boiled (100°C) for approximately 1 min anddispensed after cooling to holding temperatures. Heat-labile supplements such as blood, egg yolk emulsion,and IsoVitaleX (BBL) were prewarmed briefly andadded to appropriate media after each medium hadbeen autoclaved and allowed to cool to the statedholding temperatures. For making chocolate medium,tryptic soy agar base was used; after being autoclaved,the mixture was enriched with an equal volume ofprewarmed (60°C) 2% hemoglobin, plus 1% IsoVitaleX(vol/vol). The same procedure was used for makingegg yolk medium, except that the mixture was en-riched with prewarmed (37°C) 10% egg yolk emulsion.Phenylethyl alcohol medium consisted of tryptic soyagar base, phenylethyl alcohol (2.5 g/liter), 5% sheepblood, and the solidifying agents.

840

on June 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

GELRITE AS AN AGAR SUBSTITUTE 841

TABLE 1. Media and other ingredientsAmt (g/liter) of:

Media and ingredients' Source LotBase Agar GELRITE Magnesium

BS BBL 1123X 32 20 10 0.75BHI Difco Laboratories, 577772 37 15 7.5

Detroit, Mich.BG BBL 1122X 38 20 9 0.75Chocolate 27 13 8 0.75EY 27 13 8 0.75EMB Acumedia 8110-103 22.5 13.5 7.5 0.75MacConkey BBL 1121X 36.5 13.5 7 0.75N Difco 495555 8 15 8 0.75PEA 27 13 8 0.75SS Acumedia 8110-100 46.5 13.5 5.5TSI Acumedia 8110-101 46.5 13.5 6 0.75TSB BBL JODG2M 27 15 7.5 0.75TS agar with 5% sheep blood GIBCO 1731241TS GELRITE with 5% sheep blood GIBCO 1731247TS agar base GIBCO 180388YE Difco 647268PEA ICN Pharmaceuticals 27411

Inc., Irvine, Calif.EY emulsion Oxoid Ltd., Basing- 50122

stoke, EnglandIsoVitaleX BBL A4HDCCAgar Difco 683803GELRITE (S-60) Kelco KlA93-63002AMgCl2 * 6 H20 Fisher Scientific Co., 706396

Pittsburgh, Pa.Defibrinated sheep blood BBL 292Bovine hemoglobin solution BBL B3ESWL

I BS, Bismuth sulfite; BHI, brain heart infusion; BG, brilliant green; EY, egg yolk; EMB, eosin methyleneblue; N, nutrient broth; PEA, phenyl ethyl alcohol; SS, salmonella-shigella; TSI, triple sugar iron; TSB,Trypticase soy broth; TS, tryptic soy; YE, yeast extract.

Viable cell counts. Each of three replicate brothtubes (10 ml per tube) was inoculated with a loopful ofculture from each of the test organisms. Trypticasesoy broth (BBL) was used with all cultures exceptStreptococcus pneumoniae, the anaerobes (Clostridi-um perfringens, Peptostreptococcus anaerobius, and

Bacteroides fragilis), and Campylobacter fetus subsp.jejuni. These were grown in brain heart infusion broth.When fastidious organisms were tested, the mediawere appropriately enriched as shown in Table 2.Except for the Campylobacter culture, which was

incubated at 42°C, all the other broth cultures were

TABLE 2. Compositions of media used in quantitation of fastidious microorganismsComposition' of:

Liquid medium Plating medium with agar or GELRITE

Haemophilus influenzae CL1826 TSB + Iso + XV TSB + 10% LHB + IsoHaemophilus parainfluenzae CL1824 TSB + Iso + XV TSB + 10% LHB + IsoStreptococcus pyogenes CL1925 TSB + 10% HS TSB + 10% HSStreptococcus agalactiae CL 1342 TSB TSB + 10% HSStreptococcus faecalis CL1776 TSB TSB + 10% HSGroup C Streptococcus species CL909 TSB TS13 + 10% HSVibrio cholerae CL2020 TSB TSB + 10% HSGroup G Streptococcus species CL1929 TSB + 10% HS TSB + 5% DSBMoraxella species CL387 TSB + 10% HS TSB + 5% DSBStreptococcus pneumoniae CL1842 BHI + 5% DSB BHI + 5% DSBCampylobacter fetus subsp. jejuni CL2312 BHI + 5% DSB TSB + 5% DSBClostridium perfringens CLA144 BHI + 5% DSB BHI + 0.5% YE + He + MePeptostreptococcus anaerobius CLA107 BHI + 5% DSB BHI + 0.5% YE + He + Me + 5% DSBBacteroides fragilis CLA177 BHI + 5% DSB BHI + 0.5% YE + He + Me

a TSB, Trypticase soy broth; Iso, 1% IsoVitaleX; XV, Taxo X and V Strips (BBL); LHB, lysed horse blood;HS, horse serum; DSB, defibrinated sheep blood; BHI, brain heart infusion; YE, yeast extract; He, hemin (5 pg/ml); Me, menadione (0.5 ,ug/ml).

VOL. 46,1983

on June 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

842 SHUNGU ET AL.

incubated at 35C. The incubation period required toachieve maximum growth varied from 18 to 24 h formost of the test organisms to 48 and 72 h for P.anaerobius and C. fetus subsp. jejuni, respectively.The cultures were grown undisturbed except for thetwo strict aerobes, Moraxella and Pseudomonas spe-cies, which were grown on a Gyrotory shaker (NewBrunswick Scientific Co., New Brunswick, N.J.) at250 rpm. Anaerobes were incubated in an anaerobicchamber (Forma Scientific, Marietta, Ohio) in anatmosphere of 10% hydrogen-8% carbon dioxide-82%nitrogen. C. fetus subsp. jejuni cultures were incubat-ed in a microaerophilic environment with GasPak jarsand CampyPak II envelopes (BBL).

After incubation, the broth cultures were dilutedserially with Trypticase soy or brain heart infusionbroth. Portions (0.1 ml) of suitable dilutions werespread with a sterile glass rod over the surfaces ofplating media containing agar or GELRITE. The plat-ing was done in duplicate or, with the more fastidiousorganisms, in triplicate. The basic plating media towhich agar or GELRITE were added varied accordingto the growth requirements of the test organisms.Except where indicated (Table 2), Trypticase soybroth with or without the additives was used through-out. Nutrient broth was used with Proteus vulgaris, P.mirabilis, and Morganella morganii to prevent swarm-ing typically observed with these organisms.

Plated cultures were incubated under the sameconditions as those described for liquid cultures. Colo-nies were counted with the aid of a Biotran II automat-ed colony counter (New Brunswick Scientific Co.).Visual or manual counting was used only in thoseinstances in which defibrinated sheep blood or lysedhorse blood was incorporated into the media.Colony morphology and biochemical reactions. In the

portion of this study which involved a comparativeevaluation of colony characteristics, hemolytic pat-terns, and biochemical reactions, the freshly preparedplates of media were inoculated by streaking forisolation, and the only tubed medium used (triple sugariron) was inoculated by the usual stabbing of the buttand streaking of the slant. All cultures were incubatedat 37°C for 24 h, except for the anaerobes and theCampylobacter isolate, which were incubated for 48 hat 37 and 42°C, respectively.

RESULTS AND DISCUSSIONQuantitative determination of viable cells. Du-

plicate or triplicate viable cell counts were madeon each of the three subcultures of each of the 50clinical isolates tested in this study, and the datashown in Tables 3 and 4 represent, for each teststrain, the mean of six or nine individual counts± standard deviation. For the 36 nonfastidiousbacterial species grown on unsupplemented me-dia (Table 3), it can be seen that the mean countson media containing GELRITE were higher thanthose on agar-containing media in 29 instances.However, the analysis of variance (4) based onindividual counts within each paired set of datashowed the higher GELRITE mean to be statis-tically significant for only 8 of the 29 cultures,namely, Citrobacter freundii, Enterobacter

APPL. ENVIRON. MICROBIOL.

aerogenes, Klebsiella oxytoca, Pseudomonasmaltophilia, Shigella dysenteriae, S. flexneri,Staphylococcus aureus, and Yersinia enteroco-litica.The differences were not statistically signifi-

cant for the seven remaining cultures (E. ag-glomerans, P. mirabilis, M. morganii, Salmo-nella enteritidis, S. typhi, P. aeruginosa, andArizona hinshawii), for which lower means wererecorded for GELRITE.As for the 14 fastidious cultures (Table 4), a

comparison of the mean counts for each testculture between GELRITE and agar showed 10of the 14 cultures to have slightly higher valueson agar than on GELRITE. For the remainingfour cultures, which included Streptococcusagalactiae, S. pneumoniae, group G Streptococ-cus species, and Peptostreptococcus anaero-bius, the mean counts were higher on GEL-RITE. A statistical evaluation of these data byanalysis of variance showed the above-men-tioned differences to be insignificant.

Overall, it is obvious that GELRITE com-pared favorably to agar in its ability to serve as agelling agent in some of the basic and nonselec-tive bacteriological media used in laboratoriesfor quantitating bacterial species with differentnutritional requirements. Because in most in-stances (33 of 50 isolates) viable bacterial countswere higher on GELRITE-containing media, itwould appear that the new product is less inhibi-tory than agar to bacterial growth under theconditions of these experiments. Blood GEL-RITE medium gelled at approximately 46 to48°C. A low holding temperature (50°C) causedthe medium to solidify during pouring, whereasa high holding temperature (60°C) caused sub-stantial hemolysis of the sheep blood. This prob-lem was somewhat overcome by using smallbatches (250 ml) of the basal medium at 60°C andadding prewarmed (37 to 40°C) blood supple-ments immediately before pouring. The relative-ly short exposure time of the sheep blood to thehigh temperature prevented or greatly mini-mized the degree of hemolysis and allowedpouring of satisfactory blood plates.Colony characteristics and biochemical reac-

tions. Blood agar and blood GELRITE plateswere used to gather information on the size,color, and surface properties of representativecolonies as well as on the pattern and the size ofthe zones of hemolysis for each of the 50 bacteri-al species tested. Data on the sizes of thecolonies and of the zones of hemolysis wereexpressed as the average diameters of five inde-pendent determinations. For approximately 90%of the isolates, the colony sizes on GELRITE-containing media differed from those on agar-containing media by s0.5 mm, as exemplified byS. aureus (Fig. 1). Differences of '1.5 mm were

on June 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

GELRITE AS AN AGAR SUBSTITUTE 843

TABLE 3. Viable cell counts: agar versus GELRITE

Mean' ± SD for:Microorganism

Acinetobacter calcoaceticus CL1658Arizona hinshawii CL2075Citrobacter diversus CL1519Citrobacterfreundii CL1663Edwardsiella tarda CL2076Enterobacter aerogenes CL1548Enterobacter agglomerans CL1387Enterobacter cloacae CL1779Escherichia coli CL1552Flavobacterium odoratum CL1435Hafnia alvei CL136Klebsiella oxytoca CL1060Klebsiella ozaenae CL904Klebsiella pneumoniae CL1697Morganella morganii CL1555Proteus mirabilis CL1772Proteus vulgaris CL1190Providencia alcalifaciens CL787Providencia rettgeri CL1192Providencia stuartii CL1443Pseudomonas aeruginosa CL1560Pseudomonasfluorescens CL1542Pseudomonas maltophilia CL2016Salmonella enteritidis CL1362Salmonella typhi CL1935Salmonella typhimurium CL1866Serratia liquefaciens CL1977Serratia marcescens CL1520Shigella boydii CL1727Shigella dysenteriae CL1726Shigella flexneri CL1719Shigella sonnei CL1705Staphylococcus aureus CL1500Staphylococcus epidermidis CL1803Staphylococcus saprophyticus CL1941Yersinia enterocolitica CL1626

Agar

57.8 t 9.2138.2 t 18.660.5 t 14.663.3 t 11.526.7 t 7.961.3 t 16.476.7 t 22.289.8 t 15.661.2 t 10.445.2 t 12.0104.0 t 26.242.7 t 9.955.0 t 7.246.8 t 12.7121.8 t 28.7152.8 t 14.7119.5 t 26.9141.3 t 28.5122.8 t 22.099.5 t 16.880.8 t 17.661.7 t 16.3118.7 t 14.488.8 t 25.750.7 t 11.6

100.5 t 15.670.3 t 13.4116.5 t 10.924.3 t 7.644.2 t 15.625.8 t 8.4

160.7 t 55.690.3 t 7.0120.2 t 78.028.7 t 5.838.3 t 8.3

GELRITE

60.3 t 14.1131.0 t 20.364.5 t 12.385.3 t 11.627.3 t 10.586.3 t 17.974.0 t 28.1100.7 t 22.465.3 t 9.356.8 ± 14.3

108.7 t 14.552.0 t 7.760.5 t 16.154.2 t 8.7

127.0 t 33.4153.0 t 16.6107.8 ± 19.9159.8 ± 18.3126.7 ± 17.1103.7 ± 22.268.7 ± 18.762.0 ± 24.9172.8 ± 23.987.3 ± 15.748.2 + 6.3111.3 t 20.580.7 ± 15.1121.7 ± 24.238.3 ± 4.181.2 ± 15.847.2 ± 14.6

212.2 ± 86.8114.3 ± 9.8132.8 ± 78.332.2 ± 16.359.2 ± 2.6

a Mean of six values. The counts shown were derived from the 10-6 dilution for all cultures except S. sonneiand K. ozaenae (10-s) and P. aeruginosa, P. fluorescens, and P. maltophilia (10-7).

noted between the two gelling agents with C.perfringens (1 mm), B. fragilis (1 mm), E. clo-acae (1 mm), Hafnia alvei (1.5 mm), and A.hinshawii (1.5 mm).The larger colonies were almost invariably

seen on agar-containing media, with the excep-tion of Moraxella species, P. mirabilis, S. dys-enteriae, and S. pneumoniae, which showedslightly larger colonies on GELRITE-containingmedia. The C. fetus subsp. jejuni strain testedshowed a difference of 4.2 mm in the averagesize of the colonies between the two products. Itwas subsequently noted that C. fetus subsp.jejuni cultures spread on blood agar plates pur-chased from GIBCO and Scott Laboratories,Inc., Fiskerville, R.I., but not on those pur-chased from BBL. In fact, the average size ofcolonies on blood agar plates prepared by BBLwas nearly equal to that recorded for bloodGELRITE plates provided by GIBCO. The rea-

son for this difference is not known at this time.Two cultures (S. faecalis and S. pneumoniae)

produced alpha-hemolytic patterns, seven cul-tures (Vibrio cholerae, C. perfringens, P. aeru-ginosa; S. pyogenes, S. agalactiae, group CStreptococcus species, and group G Streptococ-cus species) produced beta-hemolytic patterns,and the rest of the test cultures produced gam-ma-hemolytic patterns on both agar-containingand GELRITE-containing blood plates. The sizeof beta-hemolytic zones ranged from 0.7 mm (S.agalactiae) to 9mm (C. perfringens) on both agarand GELRITE plates. The only significant dif-ference in zone sizes of beta-hemolysis observedbetween the two gelling agents was seen with S.pyogenes (1.8 mm versus 1.0 mm), the largerzone diameter being recorded for blood agar.Further evaluation of GELRITE by compari-

son with agar was carried out with a selectedpanel of media commonly used in the isolation and

VOL. 46, 1983

on June 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

TABLE 4. Viable cell counts of fastidious strains for agar versus GELRITEMeana' SD for:

MicroorganismAgar GELRITE

Bacteroidesfragilis CLA177 72.0 ± 8.6 69.4 ± 22.1Campylobacter fetus subsp. jejuni CL2312 49.1 ± 41.0 46.3 ± 34.6Clostridium perfringens CLA144 69.4 ± 21.6 67.3 ± 17.0Haemophilus influenzae CL1826 58.0 ± 16.2 53.3 ± 8.9Haemophilus parainfluenzae CL1824 42.4 ± 6.6 37.9 ± 14.3Moraxella species CL387 198.5 ± 54.8 181.7 ± 46.2Peptostreptococcus anaerobius CLA107 83.7 ± 29.6 84.8 ± 34.1Streptococcus agalactiae CL1342 142.1 ± 37.4 158.1 ± 27.1Streptococcus faecalis CL1776 194.0 ± 58.7 185.9 ± 48.5Group C Streptococcus species CL909 226.7 ± 48.2 202.6 ± 44.7Group G Streptococcus species CL1929 96.7 ± 71.2 113.7 ± 70.9Streptococcus pneumoniae CL1842 74.7 ± 34.6 87.7 ± 36.7Streptococcus pyogenes CL1925 56.2 ± 43.2 49.8 ± 23.1Vibrio cholerae CL2020 80.0 ± 67.9 79.7 ± 49.6

a Mean of nine values except for S. pneumoniae, group G Streptococcus species, Moraxella species, and V.cholerae (mean of six values). The counts shown were derived from the 106 dilution for all cultures except S.pneumoniae, group C Streptococcus species, group G Streptococcus species, and C. fetus subsp. jejuni (10-5),V. cholerae (10-4), and B. fragilis (10-v).

biochemical characterization of clinical isolates,namely, MacConkey, triple sugar iron, brilliantgreen, bismuth sulfite, eosin methylene blue,

FIG. 1. S. aureus colonies on tryptic soy agar (A)and tryptic soy GELRITE (G) with 5% sheep blood.Plates were incubated at 37°C for 24 h.

salmonella-shigella, egg yolk, and phenylethylalcohol. Virtually all of the biochemical reac-tions and colony characteristics on all of the testmedia listed above were the same for agar andGELRITE (Fig. 2), with the only significantexception being lack of swarming by P. vulgarisand P. mirabilis on MacConkey GELRITE me-dium. In an earlier experiment in which 0.2%NaCl was used to promote solidification of aGELRITE-nutrient broth mixture, the degree ofswarming of P. mirabilis and P. vulgaris onGELRITE was equivalent to that seen on nutri-ent agar with 0.2% NaCl added. The substitutionof NaCl with 0.05% MgCl2 caused gelling ofGELRITE while minimizing swarming of theseProteus species. As it is known that electrolyte-deficient media normally do not support swarm-ing of Proteus species (3, 5), it is possible thatsubstitution of the salts sufficiently lowered thetotal electrolyte content of the medium. Howev-er, the addition of GELRITE and MgCl2 toMacConkey medium containing 0.5% NaCl pre-vented swarming of both Proteus species, whichnormally spread on MacConkey agar. The addi-tion of MgCI2 was probably the primary reasonfor the observed lack of swarming, and it mayhave acted in concert with the bile salts compo-nent of this medium to prevent the swarmingphenomenon.As was the case with blood plates, bacterial

colonies were somewhat smaller on the differen-tial and selective media when GELRITE wasused as the solidifying agent. From a technicalstandpoint, some difficulty was experienced instreaking the relatively softer surface of bloodGELRITE media, and we were not properlyequipped to rule out a possible difference in gelstrengths between agar-containing and GEL-

844 SHUNGU ET AL. APPL. ENVIRON. MICROBIOL.

on June 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

GELRITE AS AN AGAR SUBSTITUTE 845

FIG. 2. S. enteritidis cultures (24 h) growing inbrilliant green medium with agar (A) and GELRITE(G) as the gelling agents.

RITE-containing blood media. The higher gel-ling point and quicker setting properties ofGEL-RITE as compared with those of agar should beadvantageous in that more plates can be pouredwithin a limited work space and stored awaywithout the worry of disturbing the gel. Ingeneral, media gelled by agar and GELRITE

appeared opaque and clear, respectively (Fig.2). In this respect, we observed that in mostinstances, it was easier to evaluate colony char-acterisitcs and biochemical reactions on GEL-RITE-containing media.

In summary, it can be stated that, based onthe results presented here, GELRITE can besubstituted for agar in many routine mediumapplications and, in some cases, GELRITE-solidified media may actually give higher viablecell recoveries than similar media solidified withagar. It is hoped that a few apparently manage-able technical problems encountered during thecourse of this study, namely, hemolysis of bloodat the elevated GELRITE setting temperatureand slight difficulties in the streaking of certainmedia (blood and chocolate), will eventually beresolved.

It should be noted that this study dealt withpossible applications of GELRITE primarily inthe field of bacteriology, especially in clinicallaboratory settings. Information on plating effi-ciencies of this gelling agent for other microorga-nisms (e.g., algae and fungi) is very limited (1) orstill lacking.

LITERATURE CITED1. Kang, K. S., G. T. Veeder, P. J. Mirrasoul, T. Kaneto, and

I. W. Cottrell. 1982. Agar-like polysaccharide produced bya Pseudomonas species: production and basic properties.Appl. Environ. Microbiol. 43:1086-1091.

2. Moorhouse, R., G. T. Colegrove, P. A. Sanford, J. Baird,and K. S. Kang. 1981. PS-60: a new gel-forming polysac-charide, p. 111-124. In D. A. Brand (ed.), Solution proper-ties of polysaccharides. American Chemical Society,Washington, D.C.

3. Naylor, P. G. D. 1964. The effect of electrolytes or carbo-hydrates in a sodium chloride deficient medium on theformation of discrete colonies of Proteus and influence ofthese substances on growth in liquid culture. J. AppI.Bacteriol. 27:422-431.

4. Snedecor, G. W. 1965. Analysis of variance, p. 237-290. InStatistical methods. The Iowa State University Press,Ames.

5. Williams, F. D., and R. H. Schwarzhoff. 1975. Nature ofthe swarming phenomenon in Proteus. Annu. Rev. Micro-biol. 32:101-122.

VOL. 46, 1983

on June 7, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from