An indoor system designed for the study of survival of airborne micro-organisms in closed conditions has been successfully modified to allow the effect of open air to be measured. It was found that the unidentified open-air factors which are toxic for many species of microbes and rapidly lost when enclosed in conventional laboratory apparatus could be retained in the system by continuous ventilation at an adequate rate. The rate required allowed examination of Escherichia coli in aerosols generated from small amounts of material because of the short periods of ventilation required for appreciable viable decay to occur.

The validity of the system was tested by comparing the survival of E. coli in true aerosols with its survival when the droplets were held on microthread. An investigation of the role of relative humidity in open-air toxicity was included.

To facilitate the isolation of anaerobes from cultures in which swarming organisms were present Roux flasks were adapted to provide anaerobic conditions. Using these flasks, five species of anaerobes were isolated from the caecum of mice resistant to colonization with Escherichia coli and other Gram-negative organisms. These species accounted for the majority of organisms seen in films of caecal material. No aerobes were isolated and the flora of these mice apparently consisted of a limited number of anaerobes representing part of the intestinal flora of conventional mice. Three of the species were clostridia and another was identified as Propionibacterium acnes. One species resembled Catenabacterium but was not identified. One of the Clostridium species could only be isolated when the medium was supplemented with yeast extract and vitamins. This species sporulated on media with ammonium phosphate as the source of nitrogen; tryptone completely suppressed sporulation.

The two methods for measuring the haemagglutinin content of an influenza virus suspension are the haemagglutinating (HA) and chick cell agglutinating (CCA) techniques and both measure the same biological activity. With the establishment of an international reference preparation for influenza virus haemagglutinin (type A), however, it seems logical to express the haemagglutinin content of influenza vaccines in international units. Accordingly a collaborative study was arranged in order to obtain agreement on the number of units to be assigned to a British reference preparation for influenza haemagglutinin. It was agreed that the preparation contains 190 i.u. per ampoule and 1 i.u. is contained in 0·0622 mg. of the dried material.

The use of the hypergeometric distribution in calculations of sample volumes for safety testing is criticized. Alternative interpretations of the concept of ‘infective ml. doses’ lead to slightly different solutions, the simplest of which, involving the exponential function, is recommended for general use.

Airborne Semliki Forest virus and T coliphages were inactivated at a considerably enhanced rate in open air compared with enclosed air. Open air exerts its maximum sterilizing activity on viruses contained in the smallest sized particles.

Both the haemagglutinin and the neuraminidase of influenza A 2 virus have undergone progressive antigenic drift since this subtype first appeared in 1957.

In 1968 a strain showing a much greater antigenic difference was isolated in Hong Kong. This variant has been responsible for epidemics throughout the world in the past 2 years.

In the second year of its prevalence a small proportion of strains isolated showed a small but definite antigenic change, related principally to the haemagglutinin component.

This new variant formed about 4 % of all strains investigated and was detected in Britain, Portugal, France and New Zealand.

Records of seven common-cold outbreaks on the island of Tristan da Cunha are compared with the corresponding time courses given by the mathematical model of Kermack & McKendrick (1927) and with an alternative model that directly involves a constant average duration of individual infection. Using computer simulation techniques the latter model is shown to be preferred and is then closely matched to the field data to obtain values for the model parameters. Consideration is then given to the intensity of epidemics predicted by the model and to the distribution of the actual epidemics relative to the theoretical epidemic threshold.

A survey of microfilaraemia among the population of Vanua Levu, Taveuni and Koro islands in northern Fiji was conducted in 1968 and 1969 as a prelude to a campaign of mass treatment with diethylcarbamazine.

The prevalences of microfilaraemia were found in the more moist conditions of Taveuni and Koro and on the windward southern side of Vanua Levu to be higher than on the drier northern side of Vanua Levu. On both sides of Vanua Levu prevalences were lower inland than near the coast.

Under apparently similar environmental conditions those of Fijian ethnic origin exhibited a higher prevalence of microfilaraemia than that shown by Indians. This ethnic difference and a difference between the prevalences in male and female Fijians are considered to be due more to higher rates of recovery from microfilaraemia in Indians and Fijian women than to diminished exposure to mosquitoes. Mathematical models have been used as an aid to the interpretation of the data, and, where appropriate, comparison has been made with the prevalence of antibodies to dengue, an arbovirus having the same vectors.

Household infections were analysed by computer techniques. Infections in large households were not proportionately higher than in small households, indicating that transmission was not intrafamilial. The clustering of infections within households, though present, was not marked. Among the occupants of outlying settlements the prevalence of microfilaraemia was relatively low indicating a lower risk of infection due to isolation.

Field experiments were made in southern England to re-examine the possibility that mosquitoes in Britain might feed on wild rabbits and hence be vectors of myxomatosis. Mosquitoes of several species were attracted to rabbits enclosed hi cylindrical traps and in a trap in which the animal was placed in a wire mesh cage. Substantial numbers of mosquitoes were also caught biting, or attempting to bite, tethered rabbits which were not in cages or traps. Evidence that mosquitoes fed on wild rabbits under natural conditions was obtained from results of precipitin tests made on blood-smears collected from mosquitoes caught resting amongst vegetation. On a few evenings mosquitoes were seen to be attracted to healthy wild rabbits and apparently attempting to feed on them. Batches of two mosquito species collected from the field were infected with myxoma virus.

It was concluded that contrary to previous beliefs mosquitoes in Britain feed to a certain extent on wild rabbits, and therefore are potential vectors of myxomatosis. No attempts were made to assess their relative importance in the transmission of the disease, which in Britain is transmitted mainly by the rabbit flea.

Surveys of arbovirus activity in Fiji were conducted over a 10-year period from December 1959 to December 1969. No arboviruses were isolated from over 200,000 mosquitoes, 9000 ticks, or 575 serum samples. Eight thousand human and 1117 bird, bat and animal sera were tested for haemagglutination-inhibiting arbovirus antibody using a variety of group A, group B and Bunyamwera group antigens. Only a small number of low-titre reactions were found among the non-human sera, but 14% of all human sera were found to contain Group B antibody. The antibody prevalence increased with increasing age, from less than 1 % for persons born since 1950, to 70% for persons born before 1900. The age differences in prevalence could be used to estimate the time and size of previous epidemics. Differences were found in antibody prevalence between the sexes, between ethnic groups and between persons from different regions. These differences could be explained in terms of climate, location and custom.

Historical and serological evidence both suggest that all the antibody detected was due to past exposure to dengue virus. The very high proportion of the population with no dengue antibody makes Fiji a high-risk area for a further dengue epidemic. Dengue virus is known to be active in the Pacific and South-East Asia.

In the course of developing a living attenuated feline infectious enteritis (panleucopaenia) vaccine, it was found that respiratory disease-infected cats newly inoculated with this vaccine spread vaccine virus to respiratory disease-infected in-contact controls. These in-contact controls were able to infect other cats with which they were placed in contact so that after five natural transmissions in this way and two oral administrations and subsequent re-isolations, reversion to virulence became evident. It is clear that before general release of a new living feline infectious enteritis vaccine, there must be satisfactory evidence that con-current infection will not affect the safety of the modified antigen.

In cats infected with feline infectious enteritis there appears to be a short period, coinciding with the onset of leucopaenia, during which they are highly infectious. It seems possible that some infected animals may become immune carriers because virus has been recovered from the small intestine of two of four cats with significant antibody titres 22–24 days after exposure to infection.

The rates of decline (half-lives) of maternally acquired antibodies of two different specificities in a group of infants were found to be highly variable, ranging from 18 to 192 days for parainfluenza type 3 antibody (54 infants) and from 15 to 251 days for influenza A2 antibody (nine infants). For antibodies of both specificities approximately 75% of the half-lives were between 15 and 60 days. With parainfluenza type 3 antibody, and possibly with influenza A 2 antibody, the half-lives were inversely proportional to the initial antibody titre of the babies' sera. This relationship could be described by a rectangular hyperbola. Babies with high antibody titres at birth lost this antibody rapidly whereas in babies with low initial titres antibody declined over a longer period.

The half-lives of parainfluenza type 3 antibody and influenza A 2 antibody were compared with that of rubella antibody in the same group of infants (previously published). Maternally acquired viral antibodies of different specificities did not necessarily decline at similar rates in any given child. In nine infants, maternally acquired antibodies of two different specificities (rubella and parainfluenza type 3) declined at significantly different rates in the same child. It is suggested that although the half-life of antibody of a given specificity is related to its concentration in the serum, it is independent of the level of serum antibodies of other specificities.

The propensity of an attenuated strain of feline infectious enteritis (panleucopaenia) virus to spread from vaccinated cats affected with intercurrent feline respiratory disease to unvaccinated in-contact cats was eradicated by further passaging of the vaccine virus in tissue culture. No virus was recovered from, and no antibody was found in the sentinel cats in contact with seven vaccinated animals. Thus, a further 27 passages of the vaccine virus in tissue culture has eliminated the spread factor.

The spread of coagulase-positive staphylococci has been studied in a modern hospital in which most of the patients were nursed in 4-bed rooms separated from a common corridor only by low dividing walls. Acquisition of new nasal strains from patients in other bedrooms was nearly as easy as from patients in the neighbouring beds. There was no indication that subdivision of this type hindered the spread of nasal strains as compared with open wards of the ‘Nightingale’ pattern.

The distribution of neutralizing antibody to Yaba virus and 1211 agent in the sera of three Asian and one African monkey species was examined.

Cynomolgus (Macaca irus), bonnet (M. radiata) and rhesus (M. mulatta) monkeys possessed antibody to Yaba virus at incidences of 19·9, 8·4 and 0%, respectively. In African green monkeys (Cercopithecus aethiops) the incidence was as high as 76·4%.

As for 1211 agent, no Asian monkeys had neutralizing antibody and 5·5% of African green monkey sera neutralized the virus.

During a filariasis survey conducted in northern Fiji in 1968–9 examinations were made for microfilaraemia, enlarged lymph nodes and elephantiasis. Analysis of the microfilarial densities at different ages and the number of anatomical sites showing lymph gland enlargement or elephantiasis have been used to provide evidence on the clustering of infections and pathogenesis.

Although there is no evidence of clustering of risk of infection, there is evidence favouring the clustering of adult filariae in individuals. Nevertheless the number of sites of lymph node enlargement do not correspond with this finding and statistical evidence suggests that lymph-node enlargement is not necessarily associated with the near presence in the body of adult filariae, whether dead or alive.

Males of Indian ethnic origin showed a higher prevalence of elephantiasis than males of Fijian ethnic origin, but women of either ethnic race showed prevalences lower than those of men.

The onset of elephantiasis at a site does not directly reflect the number of infections sustained in the local area, but it appears that filariasis first induces for a limited period a proneness to elephantiasis. During this period a random and discrete event may induce the onset of elephantiasis. The nature of the event is unknown, but it probably is not trauma.

Studies on infections in a hospital for diseases of the skin are described. Patients were shown to acquire staphylococci in the groin and on the chest at about the same rate as in the nose. In contrast to surgical wards, many staphylococci were resistant to tetracycline but sensitive to penicillin. Nevertheless, much of the epidemic spread of staphylococci was with typical surgical-ward strains rather than with phage group II strains which might be thought typical of skin diseases.

Rapid spread of methicillin-resistant staphylococci (MeRS) in a children's hospital is described. Within 4 months of the first isolation MeRS had been isolated from infections in all clinical units. MeRS were also regularly isolated at the out-patient department. Protective isolation of one of the clinical units had no effect on the infection rate by MeRS. The use of antiseptics (Hexachlorophene and chlorhexidine) and gentamicin nose cream in children and staff members in three out of five clinical units resulted in a significant reduction of the prevalence of nose colonization by MeRS in children. In staff members a non-significant reduction of the prevalence of colonization and a significant reduction of the acquisition of MeRS was found. After a few months the infection rate decreased to zero in the units where the measures were introduced. It remained unchanged in the other units. Phage typing of two sets of strains collected at an interval of 6 months showed that the infections were mainly caused by two endemic strains of MeRS. The majority of the infections caused by MeRS was of minor importance. In 16 % of the infections a strain was isolated repeatedly and for more than 1 week. After the introduction of antiseptics a relative increase of infections by Gram-negative bacteria was observed. The significance of this phenomenon is discussed.

Sera were examined from 50 patients on the renal transplant unit, Cambridge, for antibody against cytomegalovirus by complement fixation and by immuno-fluorescence for IgG and IgM antibodies.

The incidence of antibody on admission was 84% with a possible further 8% so that nearly all had been infected at some time by CMV.

43 (86%) patients showed evidence of active infection after admission, 39 by serology and four only from the examination of post-mortem material.

Twenty-one patients produced IgM antibody and production was prolonged for years in patients that survived. Antibody production was related both to transplantation and admission to hospital.

The evidence indicated that primary CMV infections were rare, that IgM anti-body production was the result of active infection and that this could be attributed to reactivation without the need to invoke re-infection as the source although this type of patient is both susceptible and exposed to re-infection.

In animals exposed to foot-and-mouth disease virus by indirect contact, virus was recovered from the blood, milk, pharynx, vagina and rectum for variable periods of time before clinical disease was apparent. Virus instilled into the mammary gland multiplied rapidly and virus concentrations greater than 107 p.f.u./ml. were recorded within 8–32 hr., depending on the virus strain and dose inoculated. Virus multiplication was accompanied by clinical signs of mastitis but the classical signs of foot-and-mouth disease did not appear for 52–117 hr. Dissemination of virus from the mammary gland occurred within 4–24 hr. and in some animals samples taken from the pharynx, mouth, nose and vagina contained virus for periods up to 97 hr. before the appearance of vesicular lesions. Virus production in the udder declined with the appearance of virus neutralizing activity in the blood and the milk but persisted in some animals for periods of 3–7 weeks. The ability of foot-and-mouth disease virus to persist in mammary tissue was confirmed by the demonstration of virus multiplication in the udders of immune animals.