Airborne transmission of Actinobacillus pleuropneumoniae was studied as the percentage of air needed to establish airborne transmission from an infected pig unit into a neighbouring non-infected pig unit. The experiment was carried out in two containers constructed as pig units, placed 1m apart and connected by pipes. By manipulating the air pressure in the two units, the amount of ventilation air transferred from the infected pigs (unit A) to the non-infected pigs (unit B) was controlled and measured. In three experiments, between 48 and 50 specific pathogen free-pigs were randomly assigned to each of the two units. In unit A, five pigs (experiment 1) or eight pigs (experiments 2 and 3) were inoculated with A. pleuropneumoniae serotype 2. In experiments 1 and 3, 10% of the air was transferred from unit A to B; in experiment 2, 70% of the air was transferred. In the non-infected unit (B), 36% of the pigs seroconverted during experiment 2 (70% air transfer), whereas none of the pigs seroconverted in experiments 1 and 3 (10% air transfer). As air transmission between closely located pig units has been estimated to be less than 2% under field conditions, these results indicate that airborne transmission of A. pleuropneumoniae serotype 2 between closely located pig units is rare.