Selamectin, a novel avermectin, was evaluated in two controlled studies (one in Beagles, one in domestic shorthaired cats) to determine an appropriate topical dose for efficacy against adult Ctenocephalides felis felis (C. felis) fleas on dogs and cats for 1 month. For each study, animals were allocated randomly to four treatments. One treatment consisted of the inert formulation ingredients (vehicle) administered as a negative control, and the other three treatments consisted of a single topical dosage of 3, 6, or 9mgkg(-1) of selamectin. In each study, selamectin was administered as a topical dose applied to the skin in a single spot at the base of the neck in front of the scapulae. Dogs and cats were infested with 100 viable unfed C. felis (50 males and 50 females) on days 4, 11, 18, and 27. Seventy-two hours (+/-2h) after each infestation, on days 7, 14, 21, and 30, a comb count to determine the number of viable fleas present on each animal was performed. Efficacy of selamectin on day 30 was used to select an appropriate dose. For dogs and cats, percentage reductions in geometric mean flea comb counts for the three selamectin treatments ranged from 94. 6 to 100% on days 7, 14, and 21, compared with the negative-control treatment. On day 30, reductions in flea comb counts were 81.5, 94.7, and 90.8% for dogs, and 79.8, 98.0, and 96.2% for cats treated with selamectin at 3, 6, or 9mgkg(-1), respectively. For day 30 flea comb counts for dogs and cats, analysis of variance showed that the three selamectin treatments resulted in significantly (P< or =0.05) lower counts than did the negative-control treatment. For dogs and cats, geometric mean flea counts for selamectin administered at a dosage of 3mgkg(-1) were significantly (P< or =0.05) higher than those for the 6 and 9mgkg(-1) treatment dosages combined. There were no significant differences in flea counts between the 6 and 9mgkg(-1) treatments. This analysis was confirmed by linear-plateau modeling. Thus, the optimal dose of selamectin for efficacy against adult fleas for both dogs and cats, as estimated by the turning point (plateau) in the dose response curve, was 6mgkg(-1).