Ecology and epidemiology of Vibrio parahaemolyticus on the coast of Togo*

Bull. World Health Organ. 1974 51 353-360Bull. Organ. mond. Sante 5 5 3 Ecology and epidemiology of Vibrio parahaemolyticus on the coast of Togo* J. BOCKEMUHL 1 & A. TRIEMER 2 Following the identification of V. parahaemolyticus as an important agent of cholera- like gastroenteritis in Togo, studies to establish the natural reservoirs of these bacteria were performed in 1972-73. Whereas V. parahaemolyticus was rarely found in waterfrom the Atlantic ocean, the lagoons along the coast proved to be important reservoirs. The incidence of halophilic vibrios was correlated with the salinity of these waters, which is subjected to seasonal changes with the highest sodium chloride concentration at the end of the dry season. Examination of fish and shellfish samples confirmed these results: V. parahaemolyticus was isolatedfrom only 0.5 % offreshly caught fish from the Atlantic ocean but occurred in 47.3 % offish and shellfish samples from the lagoons. A still higher occurrence was encountered in specimens purchased in the markets, with 14.3 % of sea fish samples and 67.7% of specimens originating from the lagoons being positive. Since lagoons and lakes with brackish water are typical of all the coastal areas between Nigeria and Ivory Coast it is likely that V. parahaemolyticus is a widespread pathogen on the coast of West Africa. Since its first identification as the etiological agent of acute human gastroenteritis (1) V. parahaemoly- ticus has become a well-documented cause of food- borne disease, especially in Japan where every year 47-75% of all outbreaks of bacterial enteritis are attributed to this bacterium (2). Recently, diarrhoeal disease due to V. parahaemolyticus was reported from India (3), the Republic of Viet-Nam (4), Aust- ralia (5), the USA and Central America (6), and England (7). In Africa, V. parahaemolyticus was first identified during the cholera epidemic in Togo (8). There the disease gained public health importance since most of the cases showed cholera-like symp- toms and hence were clinically confused with cases of El Tor cholera (9). Whereas cholera vibrios are strictly adapted to man, V. parahaemolyticus is widespread in nature. Therefore it seemed appropriate to investigate the natural reservoirs of these newly identified patho- * From the National Institute of Hygiene, Lom6, Togo. Requests for reprints should be addressed to Dr J. Bocke- muihl, Institute of Hygiene and Microbiology, University of Wurzburg, Josef-Schneider-Strasse 2, 87 Wurzburg, Federal Republic of Germany. 1 Formerly Medecin-Chef, Department of Bacteriology and Serology. 2Laboratory Technician, formerly Department of Bac- teriology and Serology. gens on the West African coast. The studies were performed in Togo between February 1972 and April 1973. MATERIALS AND METHODS Examination of water samples Water specimens from the Atlantic ocean were examined by spreading 0.1 ml or 0.5 ml of water on plates of thiosulfate citrate bile sucrose agar. Enrich- ment cultures of ocean water had to be discontinued because of a massive overgrowth by the nonpatho- genic, halophilic bacterium V. alginolyticus. Water specimens from Lake Togo and the lagoons were investigated by the same methods from Febru- ary to May 1972. From June 1972, additional enrich- ment cultures were performed in double strength glucose salt Teepol broth.G The incidence of V. parahaemolyticus was estimated quantitatively by determining the minimum amount of water in which the vibrios were detectable. For this reason each water sample was examined simultaneously in quan- tities of 0.1, 50, and 500 ml. a US Food & Drug Administration. Isolation and identi- fication of Vibrio parahaemolyticus. In: Bacteriological analytical manual. Washington, DC, 1972 (mimeographed). 3277 - 353- 354 J. BOCKEMUHL & A. TRIEMER In order to estimate the relation between the salinity of the water and the number of V. para- haemolyticus the chloride concentration of lagoon water samples was determined. This was done by titration with silver nitrate according to the standard methods of the Deutsche Normenausschuss (10). Examination offish and shellfish Freshly caught fish from the Atlantic ocean were examined by swabbing the gills and one side of the specimen. The swabs were taken on the fishing boats and placed in tubes with 0.25 ml of sterile 3 %Y saline. Each swab was later inoculated into 8 ml each of glucose salt Teepol broth and salt colistine broth., Freshly caught fish and shellfish samples from the lagoons as well as specimens from the markets were pounded in a mortar, and amounts weighing about a US Food & Drug Administration. Isolation and identi- fication of Vibrio parahaemolyticus. In: Bacteriological analytical manual. Washington, DC, 1972 (mimeographed). 50 g were each added to 300 ml of glucose salt Teepol broth. After 18 h of incubation at 37°C the enrichment cultures were streaked on citrate bile sucrose agar. The plates were read after 6 and 18 h of incubation and suspected colonies were isolated either on citrate bile sucrose agar or on nutrient agar with 3% sodium chloride. Final identification of V. parahaemolyticus was performed as described previously (8). The Kanagawa test of haemolysis was performed on Wagatuma agar (Eiken) with 5% human erythrocytes added. The isolated strains were serotyped by slide agglutination with commercial K- antisera (Toshiba, Japan). The identification of fish and shellfish species was effected in cooperation with the Department of Fisheries, Lome, according to the methods of Irvine (11), Daguet (12), and Crosnier et al.b Fur- b Crosnier, A. et al. Fonds de peche le long des c6tes des R6publiques du Dahomey et du Togo. ORSTOM provi- sional document No. 273-SR, 1964. 7 - lOg/litre NaCI LZJ)3 15 - 20g/litre NaCl ___ > 20g/litre NaCl . Sampling points 7.* -- .~~~0.1 Occurrence of V.parahaemolyticus -in 0.1 ml'water 0.1 - _,,O1~~~~~~~~ ANECHO -- ~~~~~~PORTO0- 1 k SEGURO t e 2 3 d s km at the end of the dry season 1971-72. VIBRIO PARAHAEMOLYTICUS IN TOGO thermore, identification of representative specimens of shrimps and swimming crabs was confirmed in the laboratory of Dr M. Fishbein, US Food and Drug Administration. RESULTS Isolation of V. parahaemolyticus from water specimens Atlantic ocean. Samples of sea water were col- lected once a month at 7 points off the Atlantic coast of Togo. Between February 1972 and April 1973 V. parahaemolyticus was found in 4 of 83 samples (5 %). All strains were negative in the Kanagawa test. Lake Togo and Lagoon of Anecho. Lake Togo was examined four times by means of an inflatable boat, three times in the dry season (Fig. 1, 2) and once in the rainy season (Fig. 3). The outlet of Lake Togo, the lagoon between Togoville and Anecho, was examined in the rainy season 1972 (Fig. 3). At the end of the dry season the water level of this lagoon is very low and samples were taken from several points off the shore. V. parahaemolyticus was cultured from 63 of 158 samples of water and 2 strains were Kanagawa-positive. The surface water temperature ranged from 26°C to 35°C. The chloride concentration in Lake Togo ranged from 4.2 to 9.0 g/litre at the end of the dry season 1971-72, corresponding to about 7.0-15.0 g sodium chloride per litre (Fig. 1). In the rainy season 1972 the salinity decreased to about 10% of the above- mentioned values: 0.36-0.95 g/litre C1- (- 0.6-1.6 g/litre sodium chloride, Fig. 3). By the end of the dry season 1972-73 the chloride concen- tration had increased again but without reaching the level of the preceding year: 3.6-6.8 g/litre Cl- (~6.0-11.0 g/litre sodium chloride, Fig. 2). Lake Togo has a low water level. Two specimens Fig. 2. Lake Togo and the lagoon of An6cho. Sodium chloride concentration and occurrence of V. parahaemolyticus at the end of the dry season 1972-73. 355 J. BOCKEMUHL & A. TRIEMER 1.0 - 1.5g/litre NaCi I - 'I 7n A a *ra t4Inr E > ~3.0g/litre NaCI_ * Sampling points .50 Occurrence of V. parahaemolyticus .in 50 ml water - XX Occurrence of V. parahaemolyticus -) .500 in 500mlwtF -~PORTO %__ -~:- - SEGURO CEA'4 / L 0, ti 1 C/~~~AA ANECHO 0 1 2 3 4 5 km Fig. 3. Lake Togo and the lagoon of An6cho. Sodium chloride concentration and occurrence of V. parahaemolyticus in the rainy season 1972. from the surface and the bottom, respectively, were taken at points where the water was more than 1 m deep in order to detect differences in salinity. Slight differences were observed only in the southern part of the lake (Table 1). The examination of the lagoon between Togoville and Anecho gave similar results. At the end of the dry season 1971-72 the chloride levels ranged from 9.0 to 12.6 g/litre, corresponding to about 15.0-21.0 g/litre sodium chloride, but decreased to 0.97-2.53 g/litre Cl- (- 1.6-4.2 g/litre sodium chlor- ide) during the rainy season 1972. By the end of the dry season 1972-73 the salt concentration had risen again to 6.7-11.6 g/litre Cl-, corresponding to about 11.0-19.0 g/litre sodium chloride. These seasonal changes in salinity were paralleled by variations in the incidence of V. parahaemolyticus in lagoon water samples. Thus, at the end of the dry season 1971-72 water samples with > 12.0 g/litre sodium chloride yielded V. parahaemolyticus by direct culture of 0.1 ml of water. In the rainy season 1972 the vibrios were detectable only in 50 ml (lagoon of Anecho) or 500 ml (Lake Togo). By the end of the dry season 1972-73 the salinity had risen again without reaching the levels of the preceding year; V. parahaemolyticus was found in 50 ml of water, but no longer by direct culture of 0.1 ml. Lagoon of Lome. The examinations were per- formed between June 1972 (rainy season) and April 1973 (end of the dry season). Water samples were taken at five points off the shore at monthly in- tervals. Formerly, the lagoon ofLome was a shallow, silted area of water with profuse plant growth. By June 1972 a sanitation programme had started and since then the water has been collected into 2 dredged basins about 2 m in depth. This has reduced the surface area by about 40% without changing the total volume of water. The examinations described were performed simul- taneously with the sanitation work. There were 356 _ _,_ _ _/_ _ _ ,_ VIBRIO PARAHAEMOLYTICUS IN TOGO Table 1. Difference in the NaCI concentration of water samples taken from the surface and the bottom of Lake Togo (Point A in Fig. 1-3) Concentration of Approximate concentra- Date of Cl- (g/litre) tion of NaCI (g/litre) Depth Season examination (m) surface bottom surface bottom March 1972 8.20 8.40 13.50 13.90 1.40 dry season April 1972 8.80 9.20 14.50 15.20 1.20 end of the dry season July 1972 0.86 0.95 1.42 1.57 1.60 rainy season March 1973 6.60 6.90 10.90 11.40 1.00 end of the dry season always slight differences in the salt concentration of the water, increasing from the east to the west of the lagoon. However, in contrast to Lake Togo and the lagoon of Anecho, marked seasonal changes in salinity as well as in the occurrence of halophilic vibrios were not observed. V. parahaemolyticus was generally detectable in 50 ml of water (Table 2). Of the total of 60 samples taken 34 proved to be positive for V. parahaemolyticus. All strains were Kanagawa-negative. Evidence of V. parahaemolyticus in lagoon water from Ghana and Dahomey. In 1972 water samples were taken in Ghana from the Atlantic coast, the lagoon of Keta, and the mouth of the Volta river. V. parahaemolyticus was isolated by direct culture (0.1 ml) from 4 of 8 samples. The vibrios were found in the ocean and in the lagoon of Keta. Water samples were also collected in Dahomey from the Atlantic coast, from the coastal lagoon, and from Lake Nokue near the lake-village of Ganvie. Table 2. NaCI concentration and occurrence of V. parahaemolyticus in water samples collected at 5 points in the lagoon of Lom6 *a ~~~~Number ofApproximate Numbr o Concentration concentration Number of positive samplesDate of Clg (g/litre) of NaCI examined 0.1 50 500 (g/litre) ~~~ml ml ml 1972 17 June 0.8-1.3 1.3-2.1 5 0 3 0 9 July + 18 July 0.8-1.2 1.3-2.0 8 0 1 1 8 August 1.4-1.7 2.3-2.8 2 0 0 2 8 September 0.9-2.0 1.5-3.3 5 0 1 2 6 October 1.7-5.1 2.8-8.4 5 0 0 0 26 October 1.1-1.9 1.8-3.1 5 3 1 0 29 November 1.3-2.3 2.1-3.8 5 1 2 1 1973 4 January 1.3-2.3 2.1-3.8 5 0 3 1 1 February 1.4-2.6 2.3-4.3 5 0 2 1 16 February 1.5-2.7 2.5-4.5 5 0 4 0 13 March 1.7-2.7 2.8-4.5 5 1 3 0 4 April 2.0-3.0 3.3-5.0 5 0 1 0 Total 60 5 21 8 357 J. BOCKEMUHL & A. TRIEMER Table 3. Isolation of V. parahaemolyticus from freshly caught fish and shellfish from Lake Togo and the lagoons Number of PositiveSpecies examinations results 'akpavi' a and 'Abuevi' (Cichlidae: Tilapia spp. and Hemichromis fasciatus) 32 13 'blolo' bottom-feeders (Bagridae: Chrysichthys spp.) 2 2 bottom-feeders (Mochocidae: Syndontis spp.) 2 2 'follevi' (Characidae: Alestes spp.) 3 3 unidentified fish 2 2 'agla' swimming crab (Ca/linectes latimanus) 7 3 'bolu' shrimps (mostly Penaeus duorarum) 7 1 total 55 26 a Local names (Mina language) are in single quotation marks. Table 4. Isolation of V. parahaemolyticus from fish and shellfish samples from the markets of Lom6 Number Pstv Species of exami- Prsitise Origin nations reut fresh 'sardines' (mixture of different small species, mostly Sardinella and Ilisha spp.) 2 0 ocean dried "sardines' 42 6 ocean cured "sardines" 23 2 ocean unidentified fish 5 0 ocean ' follevi 'a (Alestes spp.) 10 9 lagoons ' blolo' (Chrysichthys spp.) 5 3 lagoons 'akpavi ' (Tilapia spp.) 12 11 lagoons ' lahub' (small pieces of dried sea- fish, used for seasoning) 14 4 mainly ocean 'kalami', fried fish 6 1 mainly ocean agla', fresh swimming crabs(C. latimanus) 37 28 lagoons dried swimming crabs 3 1 lagoons ' bolu', raw shrimps (mostly Penaeus spp.) 37 30 lagoons cured shrimps 25 4 lagoons oysters 3 0 lagoons total 224 99 a Local names (Mina language) are in single quotation marks. V. parahaemolyticus was detected by direct culture in the lagoon and in Lake Nokue (4 of 12 samples). Isolation of V. parahaemolyticus from fish and shellfish Freshly caught fish of the Atlantic ocean. Swabs were taken during 2 cruises aboard fishing trawlers of the Department of Fisheries, Lome, in 1973. The fish had been caught with dragnets in 38-46 m of water off the coasts of Togo and Ghana. The examined specimens included groupers (Epinephelus spp.), horse mackerel (Trachurus trecae), jacks (Caranx spp.), snappers (Lutjanus spp.), grunts (Pomadasys spp.), Brachideuterus auritus, bream (Dentex, Pagrus and Pagellus spp.), Pseudopenaeus spp., file-fish (Balistes spp.), and sole (Cynoglossus spp.). Only 1 of 206 specimens was found to be positive for V. parahaemolyticus. Freshly caught fish and shellfish of Lake Togo and the lagoons. Between June 1972 and April 1973, 55 specimens of freshly caught fish and shellfish from Lake Togo and the lagoons were examined. V. para- haemolyticus was isolated from 26 specimens (473 %). All strains were Kanagawa-negative. The results are summarized in Table 3. Fish and shellfish from the markets of Lome. Between June 1972 and April 1973, 224 specimens of fish and shellfish originating from 4 markets in Lome were examined. V. parahaemolyticus was isolated from 99 specimens (44.20%). One strain proved to be Kanagawa-positive. The results are presented in Table 4. Some 133 specimens belonged to species from the lagoon system; of these 86 (67.7%) were positive. However, of the remaining 91 specimens from the Atlantic ocean only 13 (14.3 %) were contaminated with V. parahaemolyticus. DISCUSSION V. parahaemolyticus is unique among enteropatho- gens in its exclusively marine ecology as well as in requiring sodium chloride for survival and growth. It is thus distinguished from cholera vibrios which survive outside the human body for only a short time. In most parts of the world polluted estuaries and coastal regions have been identified as natural habitats of V. parahaemolyticus (13-16). On the coast of West Africa the lagoon system proved to be the most important reservoir. In Togo the incidence of V. parahaemolyticus in 358 VIBRIO PARAHAEMOLYTICUS IN TOGO lagoon waters was shown to be related to seasonal changes in salinity. In the rainy season Lake Togo and the lagoons are supplied with fresh water by precipitation as well as from small brooks and streams, which mostly dry up in the hot months. The lagoons have no natural communication with the ocean. If necessary, a canal is dug near Anecho in the rainy season to avoid flood damage. In 1971 such a canal existed from 25 August to 4 September, and in 1972 an overflow was established on 4 July and lasted, with interruptions, until the beginning of September. Since in the rainy season the Atlantic ocean is sometimes rough, sea water may at times be forced into the lagoon system. This, however, seems not to cause important alterations in salinity (Fig. 3). The higher salt concentration during the dry season may be due either to the salt content of the material forming the bed of the lagoon or to evapo- ration. The alluvial zone between the ocean and the lagoons on the mainland of Togo and neighbouring countries consists of sand and, especially in the eastern part of Togo, of an alkaline, argillaceous material containing sodium chloride (M. Lamou- roux, unpublished data, 1961). Lake Togo and the lagoon of Anecho are characterized by a large surface and very low water levels, and in the dry season have no fresh water supply. Thus evapora- tion, which in Lome has been found to be at least 4 mm per day during the hot months (J. Colombani, unpublished observations, 1962), may contribute to the greatly increased salt concentration in the dry season. The sanitation of the lagoon of Lome greatly reduced the surface area without changing the total water volume. Confining the water to deep, dredged basins seems to have produced stable conditions, i.e., fresh water with negligible seasonal change in sal- inity (Table 2). Despite the low salt concentration, V. parahaemolyticus generally remained detectable in 50 ml of water. These observations may not be definitive since the studies described were performed at the same time as the sanitation works. A reassess- ment some years after stabilization of the new ecological situation in the lagoon of Lome would be more informative. In agreement with the findings of the studies on water samples the occurrence of V. parahaemolyticus was found to be much higher on freshly caught fish and shellfish samples from the lagoons (47%) than on specimens from the ocean (0.5 %). The more frequent occurrence in specimens from the markets of Lome (64.7% and 14.3 %, respectively) can partly be explained by an increase in the number of vibrios at elevated tropical temperatures, since most pro- ducts are stored without refrigeration. However, a secondary contamination by food-sellers as well as by the use of contaminated utensils seems more likely. Otherwise the isolation of V. parahaemolyticus from dried, cured, or even fried fish would be difficult to understand. Strains of V. parahaemolyticus producing a ther- mostable haemolysin, the so-called Kanagawa phenomenon (17), are considered to be primarily pathogenic (18-20). In Togo such strains were fre- quently found in patients with acute gastroenteritis (92.60%) (9), but rarely in environmental specimens (2 of 117 isolates from water and 1 of 134 strains from fish and shellfish specimens). These results agree well with reports from Japan, where V. parahaemoly- ticus is the most important agent of bacterial en- teritis (18). Lagoons and lakes with brackish water are typical of the coastal areas between Nigeria and Ivory Coast. It seems likely that V. parahaemolyticus is a widespread enteropathogen on the coast of West Africa. ACKNOWLEDGEMENTS The authors thank Mr Kuwadah, Director, and Mr Lawson, Deputy Director, Department of Fisheries, Lome, and Captain Jacobsen, Captain Jerchau, and Captain Lundt of that Department, for their valuable collaboration. Mr Delcambre, Section of Hydrology, ORSTOM, Lome, gave advice on the chemical analysis of water samples and checked some of the results reported here. Grateful acknowledgements are also made to Dr Fishbein, FDA, Washington, for advice and informa- tion, and to Mr Kpelevi, Lome, for technical assistance. This study was supported by a grant from the Deutsche Forschungsgemeinschaft. 359 360 J. BOCKEMUHL & A. TRIEMER RESUME E,COLOGIE ET EPIDtMIOLOGIE DE VIBR!O PARAHAEMOLYTICUS SUR LA COTE DU TOGO L'identification de V. parahaemolyticus comme agent responsable de gastro-enterites chol6riformes pendant l'6pidemie de cholera au Togo (1970-1972) a entraine des enquetes relatives aux reservoirs naturels de ces vibrions sur la c6te du Togo. Ces enquetes, effectuees en 1972-1973, ont porte' sur l'eau de l'oc6an Atlantique et l'eau du lac Togo et des lagunes. On a aussi examine des poissons et des crustaces peches dans ces eaux ainsi que des echantillons achetes aux march6s de Lome. Dans l'eau de l'ocean Atlantique, on n'a trouve que 4 souches de V. parahaemolyticus sur 83 prelevements (5 Y.). Par contre, 1'eau du lac Togo et de la lagune d'Anecho est caracteris&e par de nets changements sai- sonniers de la salinite et, en consequence, de la fr6quence des vibrions halophiles. Pendant la saison seche 1971/72, V. parahaemolyticus a pu etre identifie par culture directe de 0,1 ml d'eau dans des echantillons contenant plus de 1,2 g% de NaCl. Durant la saison pluvieuse 1972, les vibrions n'ont ete retrouves qu'apres enrichissement de 50 ml (lagune d'An6cho) ou 500 ml d'eau (lac Togo). A la fin de la saison seche 1972/73, les germes etaient en gen6ral decelables dans 50 ml d'eau. L'eau de la lagune de Lome ne revelait plus de diffe- rences importantes de la salinit6 entre la saison des pluies 1972 et la saison seche 1972/73, en raison de l'assainissement de la lagune qui a et6 effectu6 au meme temps que les examens rapportes ici. V. parahaemolyticus restait decelable apres enrichissement de 50 ml d'eau. En outre, les vibrions ont ete isoles d'un nombre limite d'echantillons d'eau lagunaire provenant du Ghana et du Dahomey. V. parahaemolyticus a et rarement trouve sur des poissons fralchement pech6s dans l'ocean Atlantique (1 sur 206 echantillons; 0,5 %), mais fr6quemment sur les poissons et crustaces pech6s dans les eaux lagunaires (26 sur 55 echantillons; 47,3 %). Par contre, un taux d'isolement eleve a Wt6 obtenu a partir de specimens achetes aux marches de Lome: 14,3% (13 sur 91 &chan- tillons) des poissons originaires de l'ocean, et 67,7% (86 sur 133 6chantillons) des poissons d'origine lagunaire etaient contamines par V. parahaemolyticus. Au Togo, les lagunes et la faune lagunaire se revelent des reservoirs importants de V. parahaemolyticus. Cepen- dant, les lagunes et les lacs lagunaires a eau saumatre sont typiques de toute la cote ouest-africaine entre le Nig6ria et la Cote d'Ivoire, et il parait vraisemblable que V. parahaemolyticus, ainsi que la gastro-ent6rite due a ces vibrions, sont tres repandus dans les regions cotieres de I'Afrique occidentale. REFERENCES 1. FujINO, T. ET AL. Medical journal of Osaka Univer- sity, 4: 299-304 (1953). 2. OKABE, S. Statistical review of food poisoning in Japan-especially that by Vibrio parahaemolyticus. In: Fujino, T. et al., ed. Proceedings of the Interna- tional Symposium on Vibrio parahaemolyticus. Tokyo, Saikon Pubi. Co., 1974, pp. 5-8. 3. SAKAZAKI, R. ET AL. Indian journal of medical research, 59: 1025-1034 (1971). 4. NEUMANN, D. A. ET AL. American journal of tropical medicine and hygiene, 21: 464466 (1972). 5. BATTEY, Y. M. ET AL. Medicaljournal ofAustralia, 1: 430-433 (1970). 6. BARKER, W. H. Lancet, 1: 551-554 (1974). 7. HOOPER, W. L. ET AL. Lancet, 1: 1100-1102 (1974). 8. BOCKEMOHL, J. ET AL. Zeitschrift fur Tropenmedizin und Parasitologie, 23: 308-315 (1972). 9. BOCKEMOHL, J. ET AL. American journal of tropical medicine and hygiene, 24: 101-104 (1975). 10. Deutsche Einheitsverfahren zur Wasser-, Abwasser-, und Schlammuntersuchung, 3rd ed. Weinheim (Bergstrasse), Verlag Chemie GmbH, 1960. 11. IRVINE, F. R. The fishes and fisheries of the Gold Coast. London, 1947. 12. DAGUET, J. Les poissons du Niger superieur. Dakar, Institut fran9ais d'Afrique noire, 1954. 13. SAITO, A. Nagasaki medical journal, 45: 146-158 (1970) (in Japanese). 14. ALDOVA, E. ET AL. Zentralblatt fur Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene, L Abt. Originale, Reihe A, 218: 176-178 (1971). 15. COLWELL, R. R. ET AL. Journal of milk and food technology, 36: 202-213 (1973). 16. KANEKO, T. & COLWELL, R. R. Journal of bacteriol- ogy, 113: 24-32 (1973). 17. MIYAMOTO, Y. ET AL. Journal of bacteriology, 100: 1147-1149 (1969). 18. SAKAZAKI, R. ET AL. Japanese journal of medical science and biology, 21: 325-331 (1968). 19. TWEDT, R. M. & BROWN, D. F. Journal of milk and food technology, 36: 129-134 (1973). 20. SAKAZAKI, R. ET AL. Japanese journal of medical science and biology, 27: 35-43 (1974).