A case-control study of nasal cancer in pet dogs was conducted to test the hypothesis that exposure to environmental tobacco smoke increases risk. Cases (n = 103) were selected from a teaching hospital during 1986-1990. Controls (n = 378) with other forms of cancer were selected from the same study base. Exposure to environmental tobacco smoke was evaluated by determining the number of smokers in the household, the packs of cigarettes smoked per day at home by each smoker, the number of years that each person smoked during the dog's lifetime, and the proportion of time spent indoors by the dog. The crude odds ratio for exposure to environmental tobacco smoke was 1.1 (95% confidence interval (CI) 0.7-1.8) and was unchanged after adjustment for confounders. Skull shape was found to exert a pronounced modifying effect; among dolichocephalic (long-nosed) dogs, the odds ratio for a smoker in the house was 2.0 (95% CI 1.0-4.1). A monotonic increase in the odds ratios across strata of total packs smoked and total indoor exposure to environmental tobacco smoke was found in this group of dogs, with risks of approximately 2.5 for the highest stratum. Conversely, all odds ratios for exposure to environmental tobacco smoke among short- and medium-length-nosed dogs were approximately 0.5. The data support an association between environmental tobacco smoke and canine nasal cancer.