The results of a detailed assessment of the atmospheric conditions when foot-and-mouth disease (FMD) virus was released from Burnside Farm, Heddon-on-the-Wall, Northumberland at the start of the 2001 epidemic in the UK are consistent with the hypothesis that the disease was spread to seven of the 12 farms in the immediate vicinity of the source by airborne virus, and airborne infection could not be ruled out for three other premises; the remaining two premises were unlikely to have been infected by airborne virus. The distances involved ranged from less than 1 km up to 9 km. One of the farms which was most probably infected by airborne virus from Burnside Farm was Prestwick Hall Farm, which is believed to have been key to the rapid spread of the disease throughout the country. In contrast, the results of detailed atmospheric modelling, based on a combination of clinical evidence from the field and laboratory experiments have shown that by assuming a relationship between the 24-hour average virus concentrations and subsequent infection, threshold infection levels were seldom reached at the farms close to Burnside Farm. However, significant short-term fluctuations in the concentration of virus can occur, and short-lived high concentrations may have increased the probability of infection and explain this discrepancy.