This paper presents the results of the first formal epidemiological study in Northern Ireland of risk factors for bovine tuberculosis (TB) associated with farm boundaries, neighbours and wildlife. The study was designed as a case-control study and 427 dairy herds were investigated between December 1993 and January 1995. The case herds had more than 30 cattle, and herds in which the source of infection had been ascribed to purchased cattle were avoided. Data on the Department of Agriculture Animal Health Computer were used in conjunction with data collected through a questionnaire to examine a range of possible risk factors, including the number and nature of the farm boundaries, the number of neighbours and their TB history, the number of hedgerows, the presence of badger setts, whether badger carcases had been found on the land, and the possible presence of deer. A follow-up telephone survey was conducted to minimise bias. The results highlighted two main associations with TB breakdowns, the presence of badgers, and contiguous neighbours who had had confirmed TB breakdowns. The estimated aetiological fraction for both associations was approximately 40 per cent, suggesting that although infected cattle may have a significant role in the transmission of TB their importance relative to the badger may have been over-estimated in the Northern Ireland TB scheme. The contribution of the badger is possibly one of several main reasons for the lack of significant progress in TB eradication despite strenuous efforts associated with all aspects of the scheme.