To compare the efficacy, safety, and duration of effect of four doses of botulinum toxin type A in the treatment of glabellar rhytids in females. Double-blind, randomized, parallel-group, dose-ranging trial followed by an open-label extension. Private dermatologic clinic. Eighty female subjects with moderate to severe wrinkles at maximum frown entered the study. The first 40 subjects completing the double-blind phase entered the open-label extension. Random administration of 10, 20, 30, or 40 U botulinum toxin type A in divided doses. Open-label trial: 30 U botulinum toxin type A at the same sites in divided doses. Trained observer and subject assessments of wrinkle severity at maximum frown and repose using the Facial Wrinkle Scale (0 = none to 3 = severe), subject satisfaction, and adverse events. Follow-up monthly for up to 1 year postinjection. Relapse rates and responder rates revealed benefits lasting 3 to 6 months or longer. Objectively, 10 U of botulinum toxin type A was significantly less effective than 20, 30, or 40 U. The relapse rate at 4 months was significantly higher in the 10 U group (83%) versus 40, 30, or 20 U (28%, 30%, and 33%, respectively). Subject satisfaction was high in all groups. Duration of effect and response rates were sustained during the open-label extension. Adverse effects were mild and infrequent. Twenty to 40 U botulinum toxin type A doses were significantly more effective at reducing glabellar lines than 10 U. Most subjects experienced benefits for 3 to 4 months; some subjects demonstrated effect for up to 12 months.