In a residential estate of 4000 persons, containing neither industry nor retail butchers shops, salmonellas were regularly found in the sewerage system. They were frequently found in the sewage of a portion of the estate housing 1000 persons. The range of serotypes found was wide and some types suggested an exotic origin. No overt salmonella infection in the estate was reported during the period of survey, although local general practitioners had been previously alerted. Overt infection due to serotypes found in the survey were, however, reported in other areas of Glamorgan. Multiple sampling points in the sewerage system and a serological technique for examining samples contaminated with multiple salmonella serotypes were essential for the technical success of the survey.

We should like to thank Prof. Scott Thomson for his advice in the preparation of this paper; Dr E. S. Anderson of the Central Enteric Reference Laboratory and Bureau, Colindale, for phage-typing the strains of S. typhimurium and S. paratyphi B; and Dr G. J. G. King of the Public Health Laboratory, Bournemouth, for identifying the serotypes isolated. We should also like to thank Mr T. R. Liddington and Mr J. H. Price for their technical assistance.

Toxoplasma antibodies were absent from the sera of 32 adult Hong Kong women and were present in only two of 35 adult males and five of 46 meat workers, an overall incidence of 6·2%. This is the lowest incidence recorded in a large tropical city and lower than the incidence reported for most areas of the world, whether rural or urban. The reasons for this are unknown. Of 31 sera from swine imported from China and slaughtered in Hong Kong, 22 (71%) were positive for toxoplasma antibodies.

The number of plaque-forming units is underestimated if plaques overlap. A simple model was developed to account for this bias, and tested by an extensive experimental analysis. It is shown that models of this type are inadequate, but can be modified to give objective methods for setting limits to the acceptable counting range, and for correcting the overlap bias of observed counts which fall within this range.

Where overlapping of plaques is a significant source of error, these methods will improve the efficiency of plaque assays and will render the statistical analysis of counting data more reliable.

1. Ninety-seven (83.6%) of 116 reptiles, comprising 70 lizards, 40 snakes, 4 tortoises and 2 crocodiles, yielded isolations of organisms in the Salmonella and/or Arizona groups.

2. The reptiles were captive or free-ranging; the former were drawn from all states of mainland Australia, while the latter were from West Australia only.

3. The relative prominence of Salmonella serotypes containing numerically high somatic antigens, the finding of new serotypes, of multiple infections, and of strains in subgenera II and III was remarked.

4. The lack of evidence of differences in the serotypes isolated from captive or wild reptiles (except for the isolation of S. typhimurium in creatures closely associated with man and his domestic fauna), and the apparent absence of a specific geographical distribution of serotypes in reptiles, lent support to the con clusion that reptiles provide a natural reservoir for Salmonella and Arizona strains in Australia. The possible spill-over to man, his domestic animals and his food stuffs is discussed.

It is a pleasure to record our indebtedness to Dr Joan Taylor for her continuous interest and support in providing confirmation and identification of many Salmonella serotypes; to Dr W. H. Ewing and later Dr R. Rhode for serotyping the Arizona strains; to Dr G. M. Storr, Curator of Reptiles in the Museum of Western Australia, for identifying the reptiles and to Dr W. S. Davidson, Commissioner of Public Health, Western Australia, for permission to publish.

The overlapping of plaques compresses their distributions and reduces observed variances. For a given distribution the reduction of the variance is substantially larger than that of the mean. To derive the error of a plaque assay from the assumption that the variance equals the mean may therefore lead to serious over-estimation of the precision of the assay.

Procedures for estimating the true error of plaque assays are developed, and their use is illustrated on experimental material.

In a survey undertaken in an intensive care unit, coliform bacilli were found to be responsible for most infections, Pseudomonas aeruginosa and Staphylococcus aureus being isolated much less frequently. Tracheostomy or endotracheal intubation predisposed to infection, but in our experience intermittent positive pressure ventilation did not significantly affect its incidence. Little cross-infection has occurred, and it has never been possible to incriminate the ventilators in its transmission.

We gratefully acknowledge the assistance we have received from colleagues in the Intensive Therapy Unit and the Department of Bacteriology during the course of this investigation.

1. An outbreak of ‘Winter Vomiting Disease’ is described involving women at a University Hall of Residence. Of the 165 women exposed, 74 vomited in the first 24 hr., and 17 more over the next 4 days.

2. Public health investigations were negative.

3. On the hypothesis that the vomiting was at least partly hysterical it was predicted that the affected would have higher N scores than the unaffected on the Eysenck Personality Inventory; that they would have a higher frequency of past attendance at the University Health Centre; and that they would be non-randomly distributed through the Hall and its annexes. All these predictions proved to be incorrect.

4. There is some slight but consistent evidence to suggest that there is a neurotic component in the small group of 28 who felt nauseated but did not vomit.

5. The histogram of time of vomiting has two peaks. It is shown that 71 out of the first 74 cases can be accounted for on the hypothesis of contamination of the food at both lunch and dinner on the first day of the outbreak, with the mean response coming 7 hr. after eating.

6. It is suggested that these results make a food-borne agent a more likely explanation of ‘Winter Vomiting Disease’ than the currently favoured airborne virus acting via the central nervous system.

The costs involved in the psychiatric side of this investigation were defrayed by a grant from the Clinical Research Committee of the Middlesex Hospital. We would like to thank the Chairman and members of the Committee for making these funds available and for their interest and confidence. We would also like to thank Dr A. W. Beard, Consultant Physician at the Middlesex Hospital for his advice and support.

1. Response to selection was achieved with all strains of Myxoma virus used.

2. Heritability of resistance to myxomatosis was determined by intra-sire correlation using survival time as an index and by an all-or-none probit analysis. Both resulted in an estimate of heritability of about 35–40%.

3. The ability of animals to survive myxomatosis varied widely with environ mental variation in time.

I am indebted to W. Menzies for his very able assistance in all aspects of the work; to Dorothy Conolly for her critical assistance with data analysis, and to Dr J. M. Rendel for suggesting the probit analysis for determining heritability.

A variety of factors that affect lecithinase activity and lecithinase production by Cl. welchii were studied. The lecithinase activity was time and temperature dependent. The optimum temperature varied from 30° to 46° C. according to the strain of Cl. welchii employed. Maximum lecithinase activity was produced after 12–60 hr. There was considerable strain variation. This could easily account for the differences in the published data.

Commercial purified lecithinase was more readily destroyed by heat than the lecithinase produced by Cl. welchii in culture media. The enzyme inactivation pattern at 60° C. was similar to that at 90° C. This is not in agreement with the reports of others who found that less enzyme was inactivated at the higher temperatures.

Acid pH values completely inactivated the enzyme. However, alkaline pH values did not significantly destroy the enzyme. The lecithovitellin reaction was completely inhibited at pH values of 1–5. The optimum pH for the reaction was around pH 7–8.

The production of lecithinase was dependent upon the pH of the culture medium. One strain produced measurable lecithinase at pH values of 5·5–8·5, whereas another strain produced lecithinase only at pH values 6–5–7–5.

Lecithin stimulated the production of lecithinase in a chemically defined medium.

A technique for neutralization tests using varicella-zoster virus propagated in primary human thyroid cells is described. The level of neutralizing antibody following chickenpox does not usually exceed a titre of 1/40 and in adults many years after infection it may be very low. After zoster a much higher and more persistent antibody response occurs. Contact with chickenpox also produced a rise in neutralizing antibody in one out of the five patients tested. One case who had had chickenpox but not zoster had a high level of neutralizing antibody and the possible reasons for this are discussed. No cross-neutralization with Herpes simplex virus was demonstrated but the rise in titre of complement-fixing antibody to HS occurring in herpetic subjects with chickenpox (Ross et al. 1965) was confirmed. Two samples of human gamma-globulin were shown to have high levels of neutralizing antibody to V-Z virus and one was known to have been found effective clinically.

We are indebted to Dr R. E. Hope-Simpson for the paired contact sera and also for many valuable discussions in the course of this work. We would also like to thank Dr D. Taylor-Robinson who collected the paired zoster sera, Dr R. M. Rawcliffe and Dr F. Whitwell who arranged the supply of thyroid tissue, and Dr G. C. Turner and Dr J. Eller for the samples of human gamma-globulin. We are grateful to Mrs M. Stewart for valuable technical assistance.

The theory of assigning an upper limit to the estimate of the degree of contamination of a medium is briefly explained. Tables to save computational labour are presented and their use elucidated by examples.

The lecithinase and haemolytic activity of thirty strains of Cl. welchii isolated from food, faeces, and soil, was studied. The strains from foods produced smaller amounts of lecithinase and were, in general, less haemolytic than the strains isolated from soil and faeces.

The haemolytic activity of Cl. welchii on erythrocytes from different animal species displayed considerable variation. Sheep erythrocytes were the most sensitive to the action of alpha toxin, whereas rabbit blood was most sensitive to haemolysis by theta toxin. The degree of haemolysis was also dependent upon the concentration of the enzyme, and temperature and length of incubation.

The haemolytic activity of commercial lecithinase was observed to be similar to the haemolytic activity of the alpha toxin of Cl. welchii. This finding provides further evidence that the haemolytic and lecithinase activities of Cl. welchii are due to one substance, the alpha toxin. Exposure of commercial lecithinase to heat resulted in the destruction of its haemolytic properties.

The antigens present in the preparation of cholera toxin consisting of the culture filtrate of 569 B strain of V. cholerae have been analysed by gel-diffusion, intra-gel absorption, and immuno-electrophoresis. The antiserum raised against the toxin was tested using the following antigens: whole cell lysates of V. cholerae 569 B and an El Tor (water) strain W 6 (proposed as a vaccine strain), suspensions of the same strains of vibrios with and without heat-denaturation, suspension of an El Tor (case) strain (GS-9) and a preparation of vibrio polysaocharide. The antitoxic serum gave four precipitin bands against the toxin preparation and nine bands against the whole cell lysate of V. cholerae. Two bands could be identified as being due to the heat-labile specific antigens present in the cholera toxin. These two bands were abolished on absorption of the antitoxic serum with a live suspension of W 6 but not on absorption with heat-denatured W 6. It has, therefore, been concluded that the proposed vaccine strain W 6 contains the specific antigens of cholera toxin. The significance of these results has been discussed in relation to the residual virulence of the proposed vaccine strain and antitoxic immunity produced in the intestinal tract of the adult rabbit following exposure to the vaccine strain.

We gratefully acknowledge the gift of purified vibrio polysaccharide from Dr C. V. N. Rao of the Indian Association for the Cultivation of Science, Jadavpur, Calcutta-32. The technical assistance of Mr B. Sanyamat and Mr I. Guhathakurta is also thankfully acknowledged.

1. High titres of neuraminidase-inhibiting antibody were detected in convalescent human sera following natural influenza A2 infections.

2. Such antibody was encountered infrequently in acute serum samples. Antibody persisted only 5–6 months after infection in the four individuals from whom serial serum specimens were available.

3. Following immunization with killed influenza virus vaccines (with adjuvant) neuraminidase inhibiting antibody was detected in human sera. The titres were in general lower than those detected in convalescent human sera.

4. The specificity of the neuraminidase-inhibiting antibody in human and animal antisera was studied. Tests with convalescent human sera using purified neuraminidase preparations and with a recombinant virus containing A2 neuraminidase and haemagglutinin distinct from that of human influenza A viruses enabled the conclusion that the antibody detected was specific for influenza A 2 neuraminidase.

We wish to thank Dr D. A. J. Tyrrell of the Clinical Research Centre, Mill Hill, London, for human serum specimens, Dr D. Breeze of Evans Medical Ltd., Liverpool, for sera from an influenza vaccine trial and Dr M. S. Pereira of the Public Health Laboratory, Colindale, London, for serial serum samples from persons who had influenza in 1957. Dr M. S. Pereira carried out the haemagglutination-inhibition tests and complement fixation tests on these sera.

We are grateful to Dr H. G. Pereira and Dr D. A. J. Tyrrell for valuable discussions during the course of this study, and to Professor C. H. Stuart-Harris for his comments on the manuscript.

Oxoid dehydrated MacConkey broth was compared with laboratory-prepared Gray's improved formate lactose glutamate medium and with two dehydrated versions of it, by participants in ten laboratories. A variety of chlorinated and unchlorinated water samples was used. The superiority of the glutamate media over MacConkey broth for the detection of coliform organisms in water was again confirmed. The minerals-modified version, either dehydrated or laboratory-prepared, gave the best results and is the medium recommended for use in water examination. Care should be taken to ensure that the final pH after sterilization is 6·7.

The identification of 18 strains of haemadsorbing viruses as parainfluenza type 4 viruses subtypes A and B is reported, as is a serological survey to show the distribution of antibody to these viruses in various age groups. The results suggest that although isolation of the virus is infrequent, infection in early childhood is common and the viruses may be important respiratory pathogens.

I wish to thank Dr G. B. Bruce White for kindly providing sera and also Dr B. E. Andrews, Dr Suzanne K. R. Clarke, Dr F. O. MacCallum, Dr E. M. Mackay Scollay, Dr Hélène J. Mair and Dr J. O'H. Tobin for sending the strains.