Sixty-one pigs were housed in an isolation unit with three compartments and five pens. Each compartment had its own ventilation system resulting in air currents flowing from compartment A (pens 1 to 3) towards compartment B (pen 4), but not towards compartment C (pen 5). Classical swine fever virus was introduced by the experimental inoculation of one pig in the middle pen (pen 2) of compartment A. The virus infected the pigs in pen 4, following the prevalent air currents, and the compartmentalisation had only a retarding effect on the transmission of the virus. The absence of infection in the pigs in pen 5, which was not different from pen 4 except for the ventilation system, indicates that the spread of virus was affected by the air currents.