The objectives of this study were first to determine the cumulative incidence of bovine spongiform encephalopathy (BSE) in the British cattle population from July 1986 to June 1997, secondly, to identify individual animal-associated risk factors that influenced the age of onset of clinical signs in confirmed BSE cases, and, thirdly, to assess the effectiveness of the measures introduced to control BSE during the epidemic. The analyses were based on the population of British cattle at risk, derived from agricultural census data collected between 1986 and 1996, and BSE case data collected up to June 30,1997. The unit of interest was individual adult cattle recorded on annual agricultural censuses between June 1986 and June 1996. Univariate and multivariate survival analysis techniques were used to characterise the age of onset of clinical signs. In total 167,366 cases of BSE were diagnosed in Great Britain up to June 30,1997. The cumulative incidence of BSE between July 1986 and June 1997 was 1.10 (95 per cent confidence interval [ci] 1.09 to 1.10) cases per 100 adult cattle at risk. Cattle from the South east, South west and Eastern regions of England had 4.26 to 5.96 (95 per cent cl 4.15 to 6.14) times as great a monthly hazard of being confirmed with BSE as cattle from Scotland. Compared with cattle born before June 1985, those born between July 1987 and June 1988 had 22.5 (95 per cent cl 22.1 to 22.8) times the monthly hazard of being confirmed with BSE, whereas those born in the 12 months after July 1988 had only 7.39 (95 per cent cl 7.24 to 7.54) times the monthly hazard of being confirmed with BSE. This reduction in hazard was directly attributable to the ban on the use of ruminant protein as a feed instituted in July 1988. Successive cohorts from 1989 to 1991 experienced further reductions in the hazard of experiencing BSE. The additional decrease in hazard observed for the 1990 cohort may be attributed to the effect of the Specified Bovine Offal ban instituted in September 1990.