A pilot study was set up for the first time in France in August 2000, to obtain more precise estimates on the BSE epidemic in France. Three categories of cattle at risk of BSE (found dead on-farm, euthanased and emergency slaughtered) were sampled exhaustively from August 7 to December 22, 2000, in the three regions assumed to be the most affected with BSE in France (Basse-Normandie, Bretagne and Pays de la Loire). The samples were checked by using Prionics tests, and positive samples were confirmed by Western blot or immunohistochemistry. The overall prevalence of positive cattle was 0.16 per cent. Multifactorial logistic regression showed that there was a significantly higher prevalence among cattle from the birth cohorts July 1993 to June 1994 and July 1994 to June 1995, than among those born before July 1993, and among the categories 'euthanased' and 'emergency slaughtered' than among the category 'dead on-farm, and a higher prevalence in the regions Pays de la Loire and Bretagne than in Basse-Normandie. No significant differences in the prevalence of BSE were observed between dairy, beef suckler and mixed herds.