Five years have passed since Campylobacter jejuni became generally recognized as a common cause of acute diarrhoea (Skirrow, 1977) – although the essential scientific discovery which led to this had been made several years earlier (Butzler et al. 1973) – so this is an appropriate time to take stock of what has happened in that time and consider what we might usefully be doing to control the disease.

A most probable number (MPN) method capable of estimating as few as ten campylobacters per 100 ml of water is described. The method gave results close to those obtained by the viable count method of Miles, Misra & Irwin (1938) with graded suspensions of Campylobacter jejuni. The method was used to test raw water samples: counts were obtained ranging from 10 to 230 campylobacters per 100 ml for 11 of 49 coastal and estuary water samples, and from 10 to 36 campylobacters per 100 ml for 7 of 44 river samples. Campylobacters were isolated from an additional 24 of the ‘negative’ samples by testing 200 ml volumes by glass microfibre filtration and enrichment culture methods. The MPN method should prove to be a useful epidemiological tool particularly suited to the enumeration of campylobacters in particulate fluids.

Dog faeces collected from a public park were cultured on selective media for Campylobacter spp. Campylobacter jejuni was isolated from 12 (4·6%) of 260 specimens. In contrast Salmonella spp. were found in only three (1·2%). Six of the 12 isolates were nalidixic acid-resistant thermophilic campylobacters (NARTC), whereas during the same period of study none were found among human isolates. Most of the campylobacter positive faeces were found during June and July. Dog faeces deposited in public places constitute only a small potential source of infection by this organism.

A major outbreak of tuberculosis occurred in cattle on a farm in Dorset between 1970 and 1976. Six hundred and twenty-six cattle were slaughtered either because they reacted to the tuberculin test or had been exposed to infection. No source of infection was found until 1974 when badgers infected with Mycobacterium bovis were first discovered.

An analysis of the tuberculin test records of this herd and the six surrounding herds indicated that tuberculosis had been a sporadic problem since the early 1960's. Two peaks of infection occurred in the most severely affected herd in 1970 and 1974 when 29·8% and 27·3% of animals, respectively, reacted to the tuberculin test. These figures are exceptionally high. During the last 20 years there have been two periods when all the herds in the area had synchronous outbreaks consistent with a common source.

Analysis indicated that cattle were at greatest risk in April and May and suggest that there was re-exposure to infection at this time each year. In addition the cattle were apparently exposed to M. bovis, at sufficiently high levels for transmission to occur, for only a relatively short period of time.

Following a major outbreak of tuberculosis in cattle on a farm in Dorset, badgers were discovered to be infected with Mycobacterium bovis. Two hundred and forty sets were found in the 1200 hectares of the study area. The sets were found predominantly in areas of Portland Sand. A high prevalence of tuberculosis was found in the badger population which was removed and repopulation prevented for 3 years. The removal of the infected badgers led to the resolution of the problem in cattle. Re-colonization of the area has progressed slowly and the cattle have remained free from infection for a period of 5 years.

A large sample of the wild mammals found on a farm in South Dorset were trapped and examined for the presence of Mycobacterium bovis following the discovery of widespread infection in cattle and badgers. M. bovis was isolated from the lymph nodes of two out of 90 rats (Rattus norvegicus) and one out of seven foxes (Vulpes vulpes) but no lesions of tuberculosis were observed. It was concluded that the badger was the only species of wild mammal which was a reservoir of M. bovis in this area.

Virulent strains of Mycobacterium tuberculosis isolated from humans are divisible into five variants by using four tests: oxygen requirement (aerobic or microaerophilic), nitrate reductase activity, susceptibility to pyrazinamide (60 μg/ml) and susceptibility to thiophene-2-carboxylic acid hydrazide (5 μg/ml). The five variants are referred to as Classical human, Asian human, bovine, African I and African II. The relation of these variants to previously described types is discussed. This simple division has been shown to be useful in epidemiological studies.

The growth of Vibrio cholerae biotype eltor and biotype classical was studied in a range of cooked foods, shellfish and raw vegetables, incubated at 22°, 30° and 37°C. Both biotypes grew in all cooked foods but growth was not demonstrated on raw shellfish. The organism multiplied on some vegetables to levels of the order of 105/g. The classical biotypes of V. cholerae showed a longer lag period than the eltor biotypes in some foods particularly when incubated at 22°C. The eltor biotypes reached a higher level in the stationary phase than the classical biotypes.

The nasal carriage rate of Staphylococcus aureus was significantly higher in hospitalized persons (children, adult females and staff) − 53·8%, – than in similar persons at a local clinic − 29·8% (P < 0·001) – in Ile-Ife, Nigeria. However, unlike studies carried out elsewhere, a higher proportion of S. aureus strains obtained from persons at the clinic were resistant to commonly used antimicrobial agents than were strains isolated in the hospital. This has been attributed to the ease at which these drugs can be obtained by the general population and used unsupervised and indiscriminately. Methicillin was the most effective antimicrobial agent against pathogenic staphylococci (2·2% resistance), followed by erythromycin (16·5% resistance), co-trimaxozole (28·0 % resistance), chloramphenicol (76·9 % resistance), tetracycline (78·6% resistance) and penicillin and ampicillin (97·8% resistance). The widespread resistance of S. aureus to penicillin and ampicillin (and other antimicrobial agents) is of clinical significance in the treatment of post-operative infections, since carriers are reportedly more prone to such infections than are non-carriers.

A purified soluble Vi antigen was used in counterimmunoelectrophoresis (CIE), passive haemagglutination (HA) and solid phase radio-immunoassay (SPRIA) for detection of serum Vi antibody. Serum Vi antibody was detected in 13 of 14 chronic typhoid carriers by both CIE and HA. SPRIA studies showed that Vi antibodies in sera of these carriers were mainly of the IgG class. Successful treatment with amoxycillin in one carrier, who initially showed a positive Vi test, resulted in a reversion in 6 months to seronegativity as measured by Vi CIE. However, Vi antibody was also detected in 10 sera (3%) from 329 control subjects by HA but not by CIE. When the CIE assay was applied to 1030 serum specimens obtained from hospitalized patients, it was able to detect one typhoid carrier, but one (0·1%) non-carrier also gave a positive Vi CIE test. It is thus concluded that Vi CIE is more specific and no less sensitive than Vi HA for the detection of the typhoid carrier state. The finding by SPRIA that the anti-Vi antibodies present in typhoid carriers' sera were mainly of the IgG class further justified the use of CIE as a precipitation test for their detection.

Antibacterial drugs (oxytetracycline, streptomycin and sulphonamides) were included in the drinking water of healthy broiler chicks from the sixth to the twentieth day of life to select a resistant gut flora. On the twenty-first day the birds were divided into three groups and reared in separate rooms until 100 days of age. One group was housed in cages with wire floors while the others were reared on litter. Faeces from adult hens were added regularly to the litter of one of these groups to determine its effect on the gut flora of the chicks.

The ecology of Escherichia coli was studied using O-serotyping, biotyping and antibacterial drug resistogram typing. The proportion of E. coli in the dominant faecal flora resistant to two to four antibacterial drugs increased with time to reach a peak several days after the drugs were withdrawn. Thereafter, the level of drug resistance in the E. coli declined equally in all three groups. The majority of organisms with multiple resistance were derived from biotypes of O-serotypes initially resistant to only one drug and were identified before the drugs were administered. The decline in the level of resistance in the dominant faecal flora after the fourth week was due to the appearance of either new O-serotypes or new biotypes of O-serotypes previously shown to be multiply resistant, and which were either sensitive or resistant to only one drug. It is probable that these new strains were derived from the food since several O-serotypes appeared simultaneously in all three groups of birds.

An investigation of the bacterial flora in over 200 homes is reported. The occurrence of potential pathogens and the levels of contamination at individual sites, particularly in the kitchen, toilet and bathroom is described and the implications for hygiene practices in the home discussed.

Trials of rodenticidal baits containing 50 p.p.m. difenacoum, 50 p.p.m. bromadiolone or 20 p.p.m. brodifacoum were carried out on farmsteads against populations of Rattus norvegicus containing difenacoum-resistant individuals. Six difenacoum treatments failed in 14–42 days of baiting. Two treatments with bromadiolone succeeded in 23 and 33 days, but four further treatments lasting 35–56 days failed to eradicate the populations. Brodifacoum gave virtually complete control of six populations in 21–73 days and of the ten residual populations left behind by the other two compounds, after baiting for a further 11–85 days. The performance of both bromadiolone and brodifacoum was well below that reported by previous investigators, indicating the possibility of low-grade resistance to these compounds in the difenacoum-resistant strain.

Paired maternal and cord sera from 100 pregnancies were tested for antibodies against herpes simplex virus, measles virus and respiratory syncytial virus by complement fixation and for antibodies against rubella virus, influenza A virus and influenza B virus by haemagglutination-inhibition. For four viruses (herpes simplex, measles, respiratory syncytial and rubella) higher levels of antibody were found in cord than in maternal sera. There was no difference between maternal and cord serum titres against influenza B virus but significantly higher levels of antibody against influenza A virus were found in maternal sera than in cord sera. This discrepancy was investigated by measuring antibodies against the surface antigens of influenza A by a complement fixation technique, and by single radial haemolysis. Both methods showed a preponderance of virus-specific antibody in cord sera. We conclude that IgG antibodies against most, if not all, viruses are concentrated on the fetal side of the circulation, but that conventional haemagglutination-inhibition techniques may fail to detect this difference.

One hundred and forty-eight sera from 101 cases of acute postnatal rubella were examined by the radial haemolysis (RH) technique currently recommended by the Public Health Laboratory Service of the United Kingdom. In 75 cases the date of onset of rash was known. No sera were RH positive until the third day after the onset of rash and one serum was still negative at 9 days. By 10 days all sera gave a zone of at least 10 mm. The antibody response detected by RH appears later than that detected by haemagglutination inhibition (HI). Thirty-nine of the 101 cases could be diagnosed by seroconversion or a significant rise in titre using the HI test (diagnosis of the remainder depending upon rubella-specific IgM tests). Using the RH test this figure rises to 48 but in ten of these cases reliance on RH would have meant a delay in diagnosis. The results also indicate that many more sera would need to be tested for rubella-specific IgM if the RH test was used instead of the HI test for evaluating possible cases and contacts of rubella.

Chlorine dioxide, bromine chloride and iodine were compared with chlorine as virucidal agents. Under optimal conditions all disinfectants were effective at low concentrations, but each disinfectant responded differently to acidity and alkalinity. Disinfection by chlorine was impaired by the presence of ammonia, but the other disinfectants retained much of their potency. Disinfection of poliovirus by iodine resulted in structural changes in the virions as seen by electron microscopy, but the other disinfectants were able to inactivate poliovirus without causing any apparent structural changes.

Levels of maternally transferred antibodies against the surface antigen of the A/Texas/1/77 strain of influenza virus showed the expected decline during infancy when measured by complement fixation (CF). However, this decline was not observed when these antibodies were measured by haemagglutination-inhibition (HI). It has been postulated that this discrepancy is due to the acquisition, in the early days of life, of non-specific serum factors which increase the HI activity of sera. The levels of these factors were determined indirectly by calculating HI: CF ratios and it was shown that the factors are rapidly acquired by children between the fifth and twentieth week of life.

Inbred strains of mice varying in susceptibility to intraperitoneal (i. p.) inoculation of HSV-1 were tested by inoculating 105 p.f.u. of the SC 16 strain into the ear pinna. Increase in ear thickness was less in resistant than in susceptible strains. In the resistant C57BL/6 strain, local replication of virus, spread to cervical ganglia and development of latent infections of the ganglia were all less than in susceptible DBA/2 mice; there was also less cellular infiltration of the inoculation site. Like resistance to i.p. inoculation, resistance to ear inoculation appears to be inherited as a dominant characteristic. The test provides a non-lethal method of distinguishing between resistance and susceptibility of inbred strains of mice to HSV-1 and may also be useful in defining resistance factors.

Spleen cells from uninfected CBA mice are more cytotoxic for Sendai virus-infected L929 cells than for uninfected cells and the lymphocytes responsible have the properties of NK cells. Preincubation of spleen cells with culture supernatants from Sendai virus-infected L929 cells increases the cytotoxicity for uninfected target cells. This increase in cytotoxicity can also be produced by pretreatment with purified mouse interferon. The enhancing effect of both the infected culture supernatants and purified interferon can be neutralized with anti-interferon serum. It is concluded that the preferential killing of Sendai virus-infected L929 cells by NK cells is dependent on the induction of interferon and that interferon will increase NK cell cytotoxicity for uninfected target cells.