Clostridium perfringens is a normal inhabitant of the human and mammalian intestine, and is also distributed in sewage and soil. Consequently, many different kinds of foods may be contaminated with the microorganisms. Welchii food poisoning is mainly caused by Type A heat-resistant (Hobbs' Type) strains of Cl. perfringens. In England and Wales, the media of infection on the food poisoning are almost invariably meat dishes made from heated meat. In Japan, about a half of the media are fish dishes made from heated fish and processed fish products. Detection of Cl. perfringens in raw fishes, which are the main material of fish dishes and of processed fish produts, had been made only by Nakagawa et al. and Zen-yoji et al. Toxicological Types, Hobbs' Serological Types and other characteristics of the strains isolated from fishes are not yet revealed. In this paper, therefore, detection of Cl. perfringens in raw fishes was undertaken for the period from June to October 1966, and the incidence of contamination with the microorganisms was made clear. At the same time, the isolates were typed toxicologically, and were tested for serological relationships to Hobbs' Types. Moreover, heat resistance and biochemical properties of the isolates were also tested and were discussed. In the first place on the investigations, a new differential enrichment-medium (LAS medium) was prepared, and detection of Cl. perfringens became easy and accurate by using the medium. Identification of isolates as Cl. perfringens also became easy and exact by applying a new sporulation medium (AGACo medium) which was prepared by the author. In addition, it was made clear that a Hobbs' Type strain forms both heat-resistant and heat-sensitive spores. Observations of dormancy were made on the spores which were synchronously formed. On the other hand, detection of Cl. perfringens in processed fish products and in ground fish meat was also made. Then minimal inhibitory concentrations (MIC) of furylfuramide (FF), tetracycline (TC) and tylosin (TT) for the growth of many Hobbs' Type strains were determined as an aid of preventing food poisoning by Cl. perfringens. Besides, the inhibitory effect of FF and the influence of storage temperatures on the growth of inoculated Cl. perfringens (a Hobbs' Type strain) were tested on Kamaboko (fish cake) and fish sausage. 1) In order to enhance selectively the growth of Cl. perfringens, a new differential enrichment-medium (LAS medium) was prepared by modifying the TSN agar for estimation of Cl. perfringens. The TSN agar was modified by additon of lactose and sodium ascorbate and by decrement of agar concentration. The LAS medium was used as an enrichment medium, and was simultaneously applied as a differential medium, that is, in the selective emrichment-culture of Cl. perfringens with the LAS medium, the presence of the microorganisms was judged by both blackening the medium and producing gas in the medium. Then the differentiation with the LAS medium was more accurate than that with TSN agar medium in which the presence of the microorganisms was judged only by blackening the medium (Table 2-1). When identified by the criteria of Benoki, in Bergey's manual (key to the species of genus Clostridium) and of Strong et al., the rate of isolation of Cl. perfringens from the 213 LAS-positive (differentiation of Cl.perfringens : positive) enrichment cultures inoculated by the body surface and alimentary canal specimens of fishes, showed the high percentages of 93, 83 and 67 respectively (Table 2-2). There was a little difference (in selecting Cl. perfringens in the LAS 46℃-cultures) between the heated and unheated specimens when the isolates from the cultures were identified by Benoki's criterion (Table 2-3), 2) A new sporulation medium (AGACo medium) was prepered by modifying Ellner's medium. The modification was made by adding ammonium ac etate and sodium ascorbate, substituting starch with glycerol, decrementing phosphate concentration and adding cobalt ion (Tables 3-1～3-5, Fig.3-1). Spores were produced in the parcentages of sporulation above from 40 to 80 by use of the AGACo medium from all tested 230 strains, regardless of Hobbs' standard strains and of Toxicological Types of test strains (Table 3-6). However, the spores formed in the AGACo medium were relatively unstable. It seemed that those spores proceeded gradually to the stage of germination during the incubation period because resistance of the spores to staining was observed to decrease along with the incubation time after the sporulation. The heat resistance of spores of Hobbs' standard 13 strains formed in the AGA medium was weak. The AGA medium appeared to be inferior to the cooked meat medium as a sporulation medium for the test of heat resistance (Table 3-7). 3) Though Cl. perfringens Hobbs' Type s train 8238 (Type 2) is a heatresistant strain, a far greater number of spors formed in the AGA, LGA or cooked meat medium were heat-sensitive (80℃, 20 min), and only a very small number of those spores were heat-resistant (Table 4-1). In the synchronized cultures with the LGA medium prepared by modifying the AGACo medium, the heat-sensitive spors rapidly proceeded from the stages of spore-formtion to the stages of germination almost without taking dormancy (Fig. 4-1 and Table 4-2). The rapidity decreased at a low temperature (3℃) or under aerobic condi tions (Tables 4-3 and 4-4). The formation process was delayed by adding sodium chloride at a concentration of 10 per cent, and both the formation process and the germination process were delayed by adding calcium chloride at a concentration of 0.95 per cent (Table 4-4). In putting together the results of those experiments and the results of the preceding observation on the spores of all tested Cl. perfringens strains including Hobbs' Type strains when the spores were form ed with the AGACo medium, it was presumed that the spores of Cl. perfringens are relatively active, and in general they have little or no significance in the dormant stage. 4) Every one of Cl. perfringens 198 strains (Benoki's criterion) isolated from fishes was Type A. For the period from June to September, Cl. perfringens was detected in the high percentages (68～88%) in the unheated specimens of the bodysurface of fishes, but the microorganisms were not detected in the specimens in October when it became cool. The rate of detection of Cl. perfringens in the alimentary canal of fishes was generally low except a few kinds of fishes (Tables 5-3 and 5-4). Though the detection rate decreased, heat-resistant (100℃, 60 min) strains were isolated selectively by applying heat-treatment (80℃, 20 min) to the specimens. And the majority of the strains isolated from the heated specimens did not agree with the identification-criterion of Strong et al. (S criterion), especially in the case of heat-resistant strains (Table 5-5). The thirty-two strains (16%) out of isolated Cl. per fringens 198 strains were in agreement with Hobbs' Types in agglutination test, that is, the 32 Hobbs' Type strains belonging to the Types 2～6, 8, 9, 12 and 13 were detected (Table 5-6). In the Hobbs' Type strains isolated from the heated specimens, the 6 strains out of 11 strains were heat-resistant. It is noticeable that heat-resistant Hobbs' Type strains were isolated thus (Tables 5-6 and 5-7). In unheated specimens, Hobbs' Type strains were detected through the period from June to September, though the detection rate showed a low percentage (Table 5-8). However, every one of the 21 strains detected as Hobbs' Type strain did not possess heat resistance (Tables 5-6 and 5-7). The above facts on the unheated specimens may be not neglected from the standpoint of food hygiene because it has been recognized that heat-resistant Cl. perfringens is not always heat-resistant in the case of cultures after the isolation, and it has been reported recently that food poisoning is caused by secondary contamination after cooking with heat-sensitive Cl. peifringens. About a quarter (48 strains) of the isolated Cl. perfringens 198 strains were salicin-fermentative (acid, or acid and gas). In heat-resistant strains, the 11 strains out of the 15 strains isolated were salicin-fermentative (Table 5-9). Then the detection rate of salicin-fermenters in the isolates from heated specimens showed a relatively high percentage of 65, but that in the isolates from unheated specimens was no more than 16 per cent (Table 5-10). Though the agreement rate of the salicin-fermenters with Hobbs' Serological Types was no more than 25 per cent, it was higher in comparison with that (13%) of the salicinunfermenters (Table 5-11). That is, there were the tendencies that the salicinfermenters agree with Hobbs' Types at higher rate than the salicin-unfermenters, and are heat-resistant. On the other hand, Cl. perfringens has been regarded as a salicin-unfermenter in the S criterion. Therefore, in the detection of Cl. perfringens in fishes as a causative agent of food poisoning, it is considered th at the identification of isolated strains by the S criterion is improper, and that the salicin-fermenters among the isolates which were in disagreement with the S criterion are especially important. Besides, there was no correlation between each of the Serological Types of Hobbs' Type strains and the salicin-fermentability (Table 5-12). 5) The detection rate of Cl. perfringens in Agekamaboko, Chikuwa and Mushikamaboko (July) was 11～27 per cent in the unheated specimens, and was 3.8～17 per cent in the heated (80℃, 20 min) specimens. The detection rate in dried fish products (July) was 27 per cent in the unheated specimens, and was 31 per cent in the heated specimens (Table 6-1). In the case of ground fish meat, the detection rate in average from June to October was 42 per cent in the unheated specimens, and was 5.1 per cent in the heated specimens (Table 6-2). Every one of the isolated Cl. perfringens 103 strains (Benoki' s criterion) was Type A. Fourteen strains (14%) out of those isolated 103 strains agreed with Hobbs' Types in agglutination test, that is, 14 strains belonging to the Type 1, 2, 4, 5, 9, 11 and 12 were detected (Table 6-4). The MIC of FF, TC and TT for the grow th of Hobbs' standard strains (13 strains) and of Hobbs' Type strains (32 strains) isolated from fishes were 7.5 ppm and less, 2.5 ppm and less and 1.0 ppm and less respectively. There was the tendency that the MIC of those antiseptics are higher in the strains isolated from fishes than in the standard strains (Table 6-6). The relationships between the legal concentration of FF in Japan and the MIC of FF are as follows. At the FF concentration of 2.5 ppm permitted in Kamaboko products, 92 per cent of the standard strains and 72 per cent of the isolated strains could not grow. At the FF concentration of 5.0 ppm permitted in sausage and ham, all the standard strains and 94 per cent of the isolated strains could not grow. In the case of fish sausage and fish ham, the inhibitory effect of FF on the growth of the test strains may be expected because the addition of 20 ppm of FF is legally permitted, and it has been estimated that the remaining activity of FF after processing is about a half or at least 30 per cent of the initial activity (Table 6-6) Besides, the MIC of FF and TT decreased in the medium with 3 per cent sodium chloride, but increased in the medium with 1.0 or 0.2 per cent glucose in some test strains. The MIC of TC was not influenced by the addition of those substances (Table 6-8). In the cased Kam aboko and fish sausage which were processed after inoculating the Hobbs' standard strain 8238, the heat-resistant spores of the inoculum survived even after the processing (treated for 55 min at 75℃ or 80℃). In the cased Kamaboko without FF, the viable count of Cl. perfringens reached 1.2 x 107 / g and more after the storage for 2 days at 30℃, and the count increased to 1.3 x 105 / g after the storage for 8 days at 20℃. In the storage at 5℃, the numbe r of the viable cells of Cl. perfringens did not increase, but the microorganisms survived also after 24 days (Table 6-10). Also in the fish sausage without FF, the viable count of Cl, perfringens came to 1.2 x 107 / g and more after the storage for 2 days at 30℃ (Table 6-11). In the cased Kamaboko with 2.5, 5.0 and 20 ppm of FF, and in the fish sausage with 20 ppm of FF, the inoculated strain could not grow even at 30℃ (Tables 6-10 and 6-11). When the standard strain 8238 was secondarily inoculated before or after cooking (boiled for 10 min) into Agekamaboko, the viable count of Cl. peyfringens in the cooked Agekamaboko reached 1.2 x 107 / g and more after one day at 30℃ without an anaerobic treatment. At 5° and 20℃, however, the growth of the inoculated strain was not recognized (Table 6-12). 6) An investigation was made on whether the strains, which were isolated from fishes and from processed fish products by using the LAS medium, agree with Benoki's criterion. As a result of the investigation, a new scheme which was illustrated in Fig. 7-1 for detection of Cl. perfringens in sea-foods by using the LAS medium has been proposed.

長崎大学水産学部研究報告, v.31, pp.1-67; 1970