The influences of drug concentration and vehicle composition on the corneal penetration of the steroid fluorometholone were studied in the albino rabbit. Aqueous dosing systems included a saturated solution and 0.1, 0.05, and 0.1% suspensions of micronized fluorometholone. Two different doses of a 0.1% oleaginous ointment were also studied. The results from the 0.1 and 0.5% suspensions show a peak aqueous humor steroid concentration at 30 min and a substantial sustaining effect with these two concentrations. The results also support the belief that moderate dilution of a suspension of a slowly soluble drug may not substantially lower the aqueous humor drug levels or, conversely, that use of a higher concentration suspension may not improve the aqueous humor drug concentration-time profile. The 0.1% suspension and the saturated solution did not produce a sustaining effect. The results demonstrate for the first time that the particles present in a dose of suspension are retained within the cul-de-sac of the eye and contribute significantly to the amount of steroid penetrating the cornea. This finding was confirmed by a study in which the eye was rinsed with saline solution 30 min after instillation of a dose of a 0.05% suspension. The rinsing procedure prematurely terminated the sustaining effect of the suspension. The results of the ointment studies show that partitioning of the lipophilic steroid from the oleaginous vehicle has a greater rate-limiting influence on corneal penetration than the dissolution rate parameter associated with the aqueous suspensions. Peak aqueous humor concentration was not achieved until 3 hr after dosing and was comparable to the 0.1 and 0.05% suspensions. Predosing of the eye with a saturated solution or the 0.1% suspension prior to dosing with ointment overcomes the inability of the ointment to provide adequate drug at short times following dosing. In this case, peak levels were achieved within 60 min and then maintained. The duration of aqueous humor levels and the amount penetrating from the ointment were greater than the suspensions, and these effects are discussed relative to the mechanism. Differences in aqueous humor levels produced by 25- and 50-mg doses of ointment were minimal. A discussion of the results from all studies is presented in the context of present theories regarding the role of the lipophilic epithelial layer of the cornea as a barrier to drug penetration.