Exposure and sensitization to dog allergen is a significant cause of asthma. In this study we investigated the distribution, aerodynamic characteristics, and particle-size distribution of the major dog allergen Can f 1. Dust samples were collected in 50 homes with a dog and 50 homes without dogs. Airborne Can f 1 concentration was measured in 28 homes with dogs and 36 homes without a dog. Particle-size distribution was determined by using 10 separate Andersen sampler measurements in a dog-handling facility, and in 10 homes with dogs, and by repeated measurements in a home with one dog. High levels of Can f 1 (> 10 μg/g) were found in dust in all but one home with a dog and in eight of 50 homes without dogs. Airborne Can f 1 levels varied greatly between the homes with dogs (range: 0.3 to 99 ng/m3). Low levels of airborne Can f 1 (range: 0.4 to 1.1 ng/m3) were detected in 11 of 36 homes without a dog. Can f 1 was predominantly associated with large particles collected on the first stage of the Andersen sampler (> 9 μm), which averaged 42 to 49% of the total allergen recovered in the dog-handling facility and in homes with dogs. Small particles (<5 μm diameter) also carried Can f 1, and these particles comprised ~ 20% of the total airborne allergen load. There was an excellent concordance between the results obtained in different sampling areas, and between the total Can f 1 recovered on the Andersen sampler and on the parallel filter. In conclusion, airborne Can f 1 was detectable in undisturbed conditions in all homes with dogs and in almost one third of the homes without dogs. In houses with dogs, a significant proportion (~20%) of airborne Can f 1 was associated with small particles (<5 μm diameter). Owing to their aerodynamic characteristics, these particles would be expected to remain airborne for a long period and, when inhaled, could penetrate into the lower airways and initiate asthma attacks.