SUMMARYThree strains of Vibrio were found to agglutinate both guinea-pig and human red cells. The vibrios differed in haemagglutination tests from fimbriated enterobacteria by showing only partial inhibition by mannose and by agglutinating guinea-pig and human red cells equally well. Electron microscopy of the vibrios revealed filamentous appendages which resembled the common fimbriae of Escherichia coli in having average diameters between 6 and 10 mp and in being more numerous on organisms from liquid media than from solid media. The fimbriae on vibrios also had some characteristics of F fimbriae of enterobacteria. In comparative studies fimbriae which resembled common fimbriae of E. coli were found on Pseudomonas multivorans and Aeromonas liquefaciens.

SUMMARYThe ability of various compounds to support swarming of Proteus was determined by making additions to a minimal medium agar on which swarming did not occur. Swarming occurred when alanine, asparagine aspartic acid, glutamic acid, glutamine, proline or serine were present, either individually or together. It did not occur when certain other amino acids were added individually to minimal medium agar but did occur when these were added together. The ability of a compound to support swarming was correlated with its ability to serve as a carbon +energy source and with the stimulation of growth rate. The swarming phenomenon is discussed in the light of these findings. I N T R O D U C T I O NMany members of the bacterial genus Proteus exhibit the phenomenon of swarming when grown on solid medium (Plate 1). When inoculated at the centre of a plate of suitable medium and incubated aerobically at 30-37" the culture grows as short organisms for a few hours; then organisms at the edge of the colony thus formed elongate and spread over the surface, i.e. they swarm. Movement stops after 1-2 hr. Division of organisms at the edge of the colony then occurs, to give short organisms, and the cycle is repeated. All organisms beyond the edge of the initial colony constitute the swarm. The swarming phenomenon of Proteus has been described by many workers, with slight variations in their descriptions (

SUMMARYSamples of milk from 1501 cows with mastitis were negative for Campylobacter jejuni. The faeces of 74 healthy Friesian cows were screened for C. jejuni: 13 % of the samples were positive during the summer when the cows were on pasture, and 51 % were positive in the winter when the cows were housed. Positive samples contained on average 1 x 104 campylobacters per g faeces.It is concluded that faecal contamination rather than udder infection is the means by which campylobacters enter milk and thereby infect man.

SUMMARYRemoval of thermophilic campylobacters from sewage at three different stages of treatment at a trickling filter sewage works has been assessed. Samples of incoming sewage, primary sedimentation effluent and final effluent were taken daily from 06.00 h to 20.00 h for 5 consecutive days and the numbers of campylobacters determined by using a most probable number method. Each sample was cultured using 2 h pre-enrichment followed by enrichment in Preston broth for 48 h and detection by plating. Over 78 % of the incoming campylobacters were removed after primary sedimentation and < 0-1 % remained in the final effluent. Campylobacter jejuni biotype I and biotype II constituted 81-5 % and 15-9 % respectively of the 232 isolates tested. Serotypes common in sewage were common in human faeces. It appears that the trickling filter sewage works removes most of the campylobacters entering the sewage works, but large numbers, estimated to be approximately 1010, are released into the environment daily from a local sewage works.

The incidence of 'thermophilic' campylobacters in foods and environmental samples has been studied over a two-year period. Of 781 environmental samples, 529 (67%) were found to contain campylobacters, and campylobacters were isolated from 835 (39%) of 2116 food samples. Sewage was almost always contaminated with campylobacters (96.6% of samples) and of the food samples both poultry (55.5%) and offal (47.0%) were commonly contaminated. Determination of the heat-stable serotypes of all strains isolated from these sources and of 921 strains isolated from human faeces showed that there was a wide distribution of serotypes in most types of sample. Serotype Pen 2 was the commonest type found in human faeces (18.9%) and it was also commonest in offal (21.3%), beef (40.0%), sewage (17.7%) and was the third commonest type in poultry. A comparison of culture media and conditions for optimal production of both cytotoxic and cytotonic enterotoxins showed that Brucella Broth incubated under microaerobic conditions for 24 h at 42 degrees C was suitable for both toxins. Detection of cytotoxic activity was most sensitive using HeLa cells. The sensitivities of two ELISA systems and a Chinese Hamster Ovary tissue culture assay for detection of cytotonic enterotoxin were comparable. Not all strains isolated from cases of enteritis in human beings produced toxin; 23.1% produced cytotonic enterotoxin and 17.5% produced cytotoxin. There was no correlation between serotype and toxin production. The wide distribution of campylobacters, indistinguishable from those isolated from cases of enteritis in human beings, leads us to conclude that simplistic statements suggesting that one particular type of food is primarily responsible for cases of human disease should not be made.