A surveillance of diarrhoeal diseases in children aged under five years was carried out from March 1986 to February 1987 in a rural area of Hebei Province, China. Using a cluster sampling method, 270 study children were selected from 30 villages to represent a population of 105,405. Parents noted episodes of diarrhoea in children on a calender using a given definition of diarrhoea. The findings were verified by trained village doctors and were recorded monthly in the county Health and Anti-epidemic Stations. Stools of diarrhoea-affected children were examined to detect causal agents of diarrhoea, once in each of four seasons of the year. Of the 270 children, 218 (80.74%) suffered from one or more than one attacks of diarrhoea with a 95% confidence interval between 75.94% and 85.54%. There was no death due to diarrhoea. The diarrhoeal incidence in children of 3-4-year age group was significantly less than that in the children of less than 3 years (p less than 0.05). In the children with diarrhoea, the number of mean episodes was 3.15 per child/year, and the total mean course of suffering was 14 days per child/year. The ratio of diarrhoeal episodes in spring, summer, autumn, and winter was about 2.5:5:2:1. July was the month of highest occurrence of diarrhoea (146; 21.28% of all episodes). The epidemiological parameters of the disease declined with the growing age of the children, and there was no gender-specific preponderance in the number of episodes. During the whole year, 21.7% of the diarrhoeal episodes were due to bacterial pathogens, 4% of the episodes were due to rotavirus, and 3.65% episodes were associated with the detection of Trichomonas hominis.(ABSTRACT TRUNCATED AT 250 WORDS)