Using a reproducible model of Candida albicans keratitis in rabbits we studied the effect of topical fluconazole, a new triazole. Candida albicans DSM 70010 (2.5X10 cells) was injected into the corneal stroma of both eyes of21 rabbits. AU eyes developed a corneal ulcer. Forty-eight hours after inoculation the animals were divided into three groups: (1) 14 eyes, received fluconazole (2 mg/ml) and the epithelium subsequently removed; (2) 14 eyes, received only fluconazole drops; (3) 14 eyes, received 0/9% NaCI: half ofthis group was also debrided. We applied one drop of either substance 10 times a day for 24 days. A further six rabbits were used to judge if the drug penetrated into the cornea and aqueous humour. There was a highly significant difference between the fluconazole groups (1, 2) and the control group (3) as to hypopyon and complications (descemetocele, corneal perforation) as well as recultivation of C. albicans from corneal tissue. The difference between the fluconazole groups with and without debridement was not significant. The drug penetrated into the cornea and aqueous humour of both uninflamed and inflamed eyes.