A 2-year study from January 1981 to December 1982 was undertaken to determine the role of viruses in the causation of diarrhoea in hospitalized children in Pune, Maharashtra State, India. The stool samples of 426 children (213 diarrhoeal and 213 non-diarrhoeal controls) were investigated by electron microscopy and ELISA for the presence of viruses. Six morphologically distinct viruses were visualized: rotavirus, coronavirus-like particles (CVLP), adenovirus, astrovirus, calicivirus and small round virus-like particles (SRV). Rotavirus was detected in 28·6% of the diarrhoea! patients and in 1·4 % of the controls. The frequency of infection with rotavirus was highest in the children aged less than 5 years. The mean age of rotavirus-positive patients was 11 months. Although rotavirus was detected in almost every month, there has a seasonal trend for colder months when CVLP cases were fewest. However, the prevalence of CVLP was greater in the control group (23–0%) rather than in those with diarrhoea (8–9%). In the control group, CVLP were detected more frequently during the summer months. An inverse relationship between CVLP and rotavirus was observed in children. Adenovirus, astrovirus, calicivirus and SRV were detected in a small proportion of children with and without clinical symptoms of gastroenteritis.

Of 32983 specimens from 307 sources in England and Wales tested in the Virus Reference Laboratory for anti-HIV between 1984 and 1987, 6491 (20%) were positive. Ninety-five per cent of the positive subjects were male and 44% of them were from three London genito-urinary medicine clinics. In 1987 the numbers of newly diagnosed HIV infections decreased in homosexual men and haemophiliacs and increased in injecting drug abusers; 148/119 (12%) of all the positive findings in 1987 were in females. Between 1984 and 1987 the proportion of anti-HIV positive individuals who were asymptomatic fell by nearly 10% and the proportion with AIDS/ARC rose by nearly 10%. Of the requests leading to positive results 1280 (20%) were recognized as duplicates of previous positive results, while for 34 % of the requests no clinical information was provided. These deficiencies in the data compromise HIV surveillance based on diagnostic testing, and supplementary bias-free data are needed.