Twenty-four beagles were randomly allocated into four groups of six and housed in separate cages. Each dog was infested with 25 Ctenocephalides canis and 25 Ctenocephalides felis felis and two days later (day 0) the dogs in groups 1, 2 and 3 received a spot-on application of selamectin (6 mg/kg), imidacloprid (10 mg/kg), or fipronil (6-7 mg/kg), respectively, while the dogs in group 4 were not treated. The dogs were combed 48 hours later, the fleas were removed, counted and their species were determined. All the dogs were reinfested with the same number of the two species of fleas on days 7, 14, 21, 28 and 35, and the efficacy of the treatments was calculated 48 hours after each infestation. The mean numbers of fleas on the control dogs were 19.8 C. canis and 14.7 C. felis felis. The three treatments were effective for the full 35 days of the trial; over the first 28 days, the efficacy of selamectin ranged from 81 to 100 and 92 to 99 per cent against C. felis felis and C canis, respectively, the efficacy of imidacloprid ranged from 98 to 100 per cent and the efficacy of fipronil was 100 per cent against both species. There were no significant differences between the three treatments.