The activity of selamectin, fipronil and imidacloprid against larval cat fleas (Ctenocephalides felis felis) was evaluated in an in vitro potency assay system. One hundred microliters of each compound at various concentrations in acetone were added to glass vials (1.5 by 3 cm) to which had been previously added 20 mg of sand and 10 mg of flea feces. Vials were then ball milled to allow the acetone to evaporate. Selamectin and fipronil were tested at 0.001, 0.003, 0.005, 0.01, 0.03, 0.05, 0.11, 0.3, and 0.5 microg of active compound per tube. Imidacloprid was tested at 0.01, 0.03, 0.05, 0.1, 0.3, 0.5, 1.0, 3.0, and 5.0 microg of active compound per tube. Thirty first instar C. felis larvae were added to each vial. The number of larvae remaining alive in each vial was determined once daily for 72 h. With selamectin, reductions of >/=93.5% were achieved at 24 h after exposure at doses of >/=0.3 microg. In contrast, at 24 h neither fipronil nor imidacloprid reached 90% reduction, even at the highest doses tested (0.5 microg for fipronil and 5.0 microg for imidacloprid). Selamectin was significantly (P</=0.05) more potent than imidacloprid and fipronil at levels >/=0.03 microg. A similar pattern of activity was observed at both 48 and 72 h, but higher percentages of larvae were killed for each of the compounds as the incubation time increased. At 72 h selamectin was significantly (P</=0.05) more potent than imidacloprid at levels of 0.01-0.1 microg and significantly (P</=0.05) more potent than fipronil at levels of 0.003-0.01 microg. Therefore, selamectin was more potent than either fipronil or imidacloprid in killing flea larvae in this in vitro assay system.