By means of the spore fluorescent antibody technique, 31 strains of Clostridium botulinum types A (18 strains), B (10 strains) and F (3 strains) were found to belong to the same homogeneous group irrespective of their toxigenic types. Some strains of this species also cross-reacted with certain strains of Clostridium sporogenes types I, II and III and Clostridium histolyticum type II. By spore antigenic analysis it was found that Clostridium parabotulinum contained two components designated L and M, the former describing species specificity; the latter was the cross-reacting component shared by some strains of Clostridium sporogenes and Clostridium histolyticum. Following this, a scheme showing the distribution of spore antigenic components among various species of Clostridium was given.

The antigenic varieties of influenza A virus isolated from 1968 to 1976 in a surveillance of a small, rather remote population were similar to those from England and Wales as a whole, despite frequent antigenic changes during the period. Household studies in the first two H3N2 influenza A epidemics found low attack rates within households, a high proportion (70%) of affected households with only one case of influenza, similar distributions of affected households in the two epidemics by the number of cases of influenza and similar distributions of the influenza cases by the day of their onset in the household outbreak. No serial interval could be demonstrated by cumulating household outbreaks. More than one minor variant was causing influenza contemporaneously in the same villages in several seasons, and different variants were on one occasion found on successive days in bedfellows. The regular occurrence of epidemics in winter was often accompanied by the disappearance of the epidemic variants and their replacement, after a virus-free interval, by new variants.

These epidemiological findings seem best interpreted on the following tentative hypothesis. Influenza A sufferers do not transmit the virus during their illness; instead it rapidly becomes latent in their tissues so that they become symptomless carrier-hosts and develop specific immunity. Next season an extraneous seasonally mediated stimulus reactivates the latent virus residues so that the carrier-host becomes briefly infectious, though symptomless. Antigenic drift occurs because particles reconstituted to be identical with the progenitor virus cannot escape the specific immunity it has provoked in the carrier host. He can shed only mutants also determined by the progenitor virus. From the assortment of mutants shed by the carrier-host, his non-immune companions select that (those) which is best fitted to survive, and it rapidly causes influenzal illness. Epidemics consist largely or entirely of such persons sick with influenza caused by reactivated virus caught from symptomless carrier-hosts.

A study was made of the extent to which frozen broilers, contaminated with indicator organisms, can cause cross-contamination in the kitchen.

In 60 kitchens a number of relevant objects were sampled during the preparation of contaminated frozen broilers. The results show that cross-contamination occurred in a high proportion of the kitchens examined. In many instances the indicator organism was still present on various objects even after rinsing, ‘clearing’ or washing up. In view of the possible risk of a cross-contamination with Salmonella spp. the importance of instructing food preparers is emphasized.

No salmonellas could be found in the sinks of the 60 kitchens examined.

All of 86 foods routinely examined for potentially pathogenic enteric bacteria were found to harbour one or more coliform species. None of the strains isolated produced heat-labile enterotoxin (LT) or showed invasive properties. The suckling mouse test indicated that one strain of Escherichia coli produced heat-stable enterotoxin (ST). Twelve incidents of suspected food poisoning were also investigated. In two of them the foods examined contained LT-producing strains of E. coli and in two there were LT-producing strains of Klebsiella pneumoniae. The counts of viable enterotoxigenic micro-organisms in these foods were 3000–30 000 E. coli/g and 50 000 to 1 million K. pneumoniae/g. The dominant symptom in all the incidents was watery diarrhoea. These seem to be the first reported cases of foodborne enterotoxigenic enteric bacteria in Europe. Though enterotoxigenic E. coli and related gram-negative enterotoxin-producing species are rare in correctly handled food in Sweden, these micro-organisms should be searched for when outbreaks of food poisoning are investigated.

A prospective study of 2983 operations in general and orthopaedic surgery during 3 years performed in four operating theatres in a modern operating suite was carried out in order to evaluate the importance of airborne infection. Weekly nose-and-throat samples were taken from the surgical staff and pre-operative samples were taken from the nose, throat, skin and perineum of the patients. The air contamination was followed by using settle plates, which showed low mean counts of total bacteria of between 9 and 15 c.f.u./m2/min, with mean counts of Staph. aureus of between 0.03 and 0.06 c.f.u./m2/min. No correlation was found between the total number of bacteria and the incidence of post-operative infections or between the amount of Staph. aureus in the air and post-operative Staph. aureus infections. It was concluded that further increases in ventilation could, at best, only marginally affect the incidence of post-operative infection.

The post-operative wound-infection rate was 9.0%. In various types of surgery, the infection rates varied from 5.3% in clean operations to 47.6% in dirty surgery. About one third of the infections were classified as moderate or severe.

Adverse patient factors, such as immunodeficiency, steroid treatment, intensive care, etc., increased the rate to 15.0%; in ‘normal’ patients it was 3.8%.

Among the bacteria isolated, gram-negative bacilli (31% of wounds), often together with other bacteria, and Staph. aureus (28%) predominated, but in 25% no specimens were taken.

Of 76 post-operative Staph. aureus infections, 32 were caused by the patients' own strains, and of the remaining 44 infections, 22, or 8% of all infections, could be traced to strains present in the air and/or the respiratory tracts of staff during the operation.

The length of pre-operative hospital stay had no influence on the carrier rate of Staph. aureus in patients. The incidence of post-operative wound infection was significantly higher in patients carrying Staph. aureus and was even higher if these bacteria were found on the skin.

Patients with wound infections stayed, on an average, 15 days longer than patients without infections. In serious infections the increase in duration of stay was > 20 days. Although infections were commoner in older patients, the average additional hospital stay of infected patients did not increase with age. If the post-operative infections studied in a concurrent retrospective study are taken into account more than 12 000 bed-days were due to post-surgical wound infections in the period studied or, in other words, some 12 beds (corresponding to 5.5% of the total) were always occupied by infected patients.

The effect of artificial air-ionization on air-borne transmission of Newcastle disease virus (NDV) infection in chickens was studied in an isolated system consisting of two side-by-side cages with solid walls and a wire-gauze roof. During a 3-week observation period more than 90% of the uninoculated indicator chickens, housed in one of the cages, contracted the virus shed to the air by the NDV-inoculated, diseased birds in the neighbouring cage. This air-borne transmission of NDV was completely prevented by increasing the ion concentration in the test room by a constant negative corona discharge above the wire-gauze roof. On the other hand, spreading of the infection within a group of chickens housed in a single cage was not affected by air ionization.

These and other results suggest that artificial air-ionization may protect animals from certain air-borne infections by interfering with microbial aerosol formation and/or by facilitating their decay.

A solid-phase radioimmunoassay (RIA) for detecting rubella virus IgA serum antibodies was developed. Purified rubella virus grown in roller cultures of Vero cells was adsorbed onto polystyrene beads. The coated beads were then incubated with dilutions of serum, and rubella IgA antibodies which attached to the virus antigen on the solid-phase were subsequently detected with 125I-labelled anti-human-alpha antibodies. The specificity of the iodinated anti-human immunoglobulins was confirmed by RIA analysis of fractions obtained by chromatography of an early convalescent serum on an agarose column. A complete separation of IgM, IgA, and IgG was observed.

A total of 144 serial serum specimens from 31 adult patients with an acute rubella infection were tested for rubella IgA antibodies, and the results were compared with the RIA IgG and IgM titres reported earlier from the same specimens. The RIA IgA response was detected in each of the 31 patients and the IgA antibodies appeared almost simultaneously with the IgG and IgM antibodies. The maximum titres, which were lower than the IgG and IgM titres, were reached in about 1 week after the onset of rash. In 6 patients out of 31 the IgA antibody response was transient and persisted approximately two months, while in the remaining 25 patients the IgA antibodies persisted throughout the study period of more than 5 months. The results obtained indicate that the presence of rubella IgA antibodies in serum is not an indication for a recent rubella infection.

By means of techniques of analyses of survival data developed for cancer trials it is possible to study aspects of the natural history of the infection of schistosomiasis on the intermediate host of transmission, the snail.

The simultaneous study of three response variables is largely based on a model of Lagakos (1976). When using this approach in the schistosomiasis setting it seems inappropriate to assume that one process, the duration of latency, follows an exponential distribution. Thus this stage is modified to follow a normal distribution and the derivatives required to obtain maximum-likelihood estimates and approximate variances of all parameters are provided.

Simple graphical tools for assessing the validity of distributional assumptions in survival data are available from industrial research. The reader's attention is drawn to a paper by Nelson (1972). The relevance and application of these methods to the current problem are described in Section 4.

In the event that the times to death of prepatent and patent snails do not follow exponential distributions as assumed in the primary model, a further modification is introduced to enable either or both to follow Weibull densities.

Lastly it is possible to adapt both the primary model of Section Three and the modified model of Section Five to allow for the inclusion of auxiliary variables or covariates. Again the required derivatives to obtain maximum likelihood estimates and approximate variances are provided.

I wish to thank Sheila Gore, Stuart Pocock and Professor Michael Healy for many useful discussions. Sheila Gore and Professor Peter Armitage provided many recommendations for the improvement of the manuscript for which I am most grateful.

Indirect haemagglutination (IHA), agglutination and complement fixation tests (CFT) for Bordetella pertussis antibodies were compared on paired sera from 52 suspected cases of whooping cough and single sera from 83 children with no recent history of whooping cough. All three tests detected serotype antibodies 1, 2 and 3, but the IHA test was the most sensitive; in seven cases it was the only test to show a rise in titre. It is recommended, particularly with vaccinated children, that the serological diagnosis of whooping cough should be based upon a rise in titre. There should be a gap of at least 2–4 weeks between serum samples, depending on the age and vaccination state of the child. The CFT appears to detect a different antibody from that detected by the other two tests, and in three cases it was the only test to show a rise in titre.

An infusion of E. coli endotoxin (10 μg) into the mammary gland of the cow 16 h before experimental infection with St. agalactiae prevented the establishment of mastitis. A clinical examination of the affected gland did not reveal signs of inflammation, all organisms were eliminated from the milk by 36 h after infection. Mammary glands not pretreated with endotoxin, and injected with the same number of viable streptococci, showed signs of severe inflammation, the milk yield was reduced and the injected pathogen could be isolated from the gland for up to 14 days when sampling was discontinued. The injection of endotoxin alone produced a rapid increase in the cell count of the milk, there were some signs of inflammation and 80% of the cells in the milk were neutrophils. The cell counts in the milk remained above normal for 7–10 days. Untreated control quarters appeared to be unaffected by the injection of E. coli endotoxin into an adjacent quarter, on the other hand the injection of St. agalactiae into the adjacent quarter on the same side of the udder produced a cell response which reached a peak after 2 days and returned to preinoculation levels by the fourth day. No inflammation was observed in these control quarters and no streptococci were isolated from the milk.

f2 bacteriophage in the presence of fetal calf serum (at a final concentration of 10%) was exposed to six commonly used disinfectants for times of 10, 20 and 30 sec. At the end of exposure times skim milk neutralized the disinfectant activity and residual virus was assayed using the plaque technique. The 6 disinfectants considered were Javex, sodium hydroxide, ethanol, Wescodyne, One Stroke Ves-Phene and Sonacide. A 0.25% (w/v) solution of sodium hydroxide and 1/50 Javex (1200 parts/106 chlorine) were the most effective of the six disinfectants considered since 105 f2 bacteriophage were inactivated in 30 seconds in each instance. Since a 0.25% (w/v) solution of sodium hydroxide had a pH of 12.5 this made it too caustic to use as a disinfectant in many practical situations. It was concluded therefore that Javex at some dilution less than 1/50 (greater than 1200 parts/106 chlorine) was the most practical of the six disinfectants to use. Ethanol (95%, v/v) inactivated 103 f2 bacteriophage in 30 seconds while 1/20 Wescodyne and undiluted Sonacide inactivated 101 virus particles. Ves-Phene at a dilution of 1/50 was a completely ineffective virucide during the 30 sec exposure. The resistance of f2 bacteriophage to inactivation by these six disinfectants was compared with that of echovirus 11 and coxsackievirus B5. In all instances except exposure to undiluted Sonacide, f2 was comparable in resistance to inactivation and in many cases had greater resistance.

A bacterial spore test has been developed which enables the efficacy of the sterilizing cycle recommended by the British Pharmaceutical Codex (1973) for bottled fluids to be accurately monitored. During a 14-month period this test detected faults in 3.3% of the sterilizing cycles, representing five distinct episodes of sterilization failure that passed unnoticed by the conventional controls of physical measurements and sterility testing. There were no failures of sterilization as detected by conventional techniques which were not indicated by the spore test.

A solid-phase micro-enzyme-labelled immunosorbent assay (ELISA) using guinea pig antiserum against purified (140S) inactivated foot-and-mouth disease (FMD) virus has been usedin a sandwich technique to specifically measure 140S virus in the presence of 12S material.

The effect of long-term exposure to cigarette smoke on the height and specificity of the secondary humoral immune response to influenza was investigated in a murine model system. It was shown that if mice were pre-immunized with a sub-lethal infection of influenza virus and then exposed to cigarette smoke daily for 36 weeks, they were able to mount a secondary immune response of normal height on subsequent challenge with the homologous virus strain. The response, however, was less specific than that elicited in control mice, with high titres of cross-reacting antibody by haemagglutination-inhibition to the following strain in the same antigenic series. Recall of antibody to the previous strain in the antigenic series was not observed in either control or smoke-exposed animals. These results serve to correct an earlier discrepancy between the murine system and human studies in which the response to influenza infection in mice was depressed by prolonged exposure to cigarette smoke, whereas in man the response of smokers did not differ significantly from that of non-smokers. This apparent discrepancy had been caused by a lack of previous experience of influenza in the mice, which had therefore mounted a primary response, compared with the secondary response observed in the human studies.

Water pollution was monitored in six Gambian village wells over a period of 8 months spanning the 5-month monomodal rains and the pre- and post-rains dry periods. Faecal coliform (FC) and faecal streptococci (FS) counts were high throughout and there was a massive increase associated with the onset of the rains, maximum counts exceeding 5 x 105/100 ml. This pattern was largely sustained throughout the rainy season. Some individual variations in patterns of pollution could be ascribed to well design, in particular lining of the shaft, but no well was protected from the seasonal increase in faecal pollution. The source of the increased pollution appeared to be a flushing in of faecal material of indeterminate or mixed human and animal origin, probably over considerable distances. Peaks of pollution not associated with rainfall episodes could have resulted from the practice of communal laundering in the near vicinity of the wells. Specific pathogens including Salmonella spp. were isolated only intermittently. Attention has been drawn to a problem complicating the standard method for assessing FC counts.

High titres of endotoxin as measured by the Limulus test were usually found in burned patients who had raised body temperatures, and were colonized with gram-negative bacteria; also some infected patients showed raised endotoxin without a raised temperature.

Patients vaccinated with an antipseudomonas vaccine rarely showed endotoxin in their plasma but occasional plasma samples from vaccinated patients had a high titre of endotoxin which appeared unrelated to infection or to a raised temperature in the patient.

An injection of influenza vaccine was offered to approximately 60 000 Postal and Telecommunications staff at the beginning of five successive winters. The sickness absence of this group, which included those who accepted the offer of vaccine as well as those who did not, was compared throughout the winter with that of a similar number of employees who were not offered vaccine. The two groups, ‘vaccinated’ and control, comprised the staff of nearly 400 Post Office units scattered throughout Great Britain, the units of the two groups being matched as far as practicable for numbers employed, type of work, region and type of location.

The proportion who accepted vaccine fell from 42% in the first year (when only 26 000 Telecommunications employees were offered vaccine) to 35% in the second year, and 25% by the fifth year.

With the exception of Telecommunications employees in 1972–73, the sickness absence rate of the group offered vaccine was less than that of the group not offered vaccine, and the difference was evident during the winter observation periods both when influenza was prevalent and when it was not. In the last four years of the study the average difference in sickness absence between the ‘vaccinated’ and control groups was 1.26 days per 100 employees per week during and 1.12 days outside the influenza periods. Moreover, the difference during the influenza periods was greater than could be expected from the acceptance rate of vaccine and the estimated attack rate of influenza. The apparent reduction in sickness absence of the group offered vaccine in comparison with the groupnot offered vaccine represented an appreciable saving in cost.

It is suggested than an annual influenza vaccination campaign in industry may produce financial benefit, but that only a proportion of the benefit is due to an improvement in health.

As mentioned in Chapter 1, the Brown Institution was staffed by a remarkable series of Superintendents, seven out of eight of whom were, or became, Fellows of the Royal Society. Much of the research work they undertook has already been referred to in preceding chapters, but it seems only fitting that something should be said about their personalities and careers. In preparing the short accounts that follow, I am greatly indebted to obituary notices in the various scientific journals, especially those in the Proceedings of the Royal Society; to the pen sketches given by Professor Sharpey-Schafer (1927) in his History of the Physiological Society;and to the Memoirs in book form of Sir John Burdon-Sanderson by his wife (1911)and of Sir Victor Horsley by Sir Stephen Paget (1919).