A ®nite dose diffusion system was employed to study cuticular penetration of 2-(1-naphthyl) [1À C]acetic acid (NAA) from simulated spray droplets through enzymatically isolated tomato fruit cuticles (Lycopersicon esculentum Mill cv Pik Red). Isolated cuticles were mounted on diffusion halfcells with the cell wall surfaces facing a 20 mM citric acid receiver solution (pH 3.2, volume 2.9 ml, prepared with deionized water). A 5-ml donor droplet containing NAA at 100 mM in 20 mM citric acid buffer (pH 3.2) was applied to the outer surface. Penetration was monitored by repeated sampling of the receiver solution. NAA penetration was characterized by (1) an initial lag phase of about 2.3 h, (2) a phase of nearly constant maximum rate of penetration averaging 6.3% of applied NAA h À1 (equivalent to 0.032 nmol h À1 ) and (3) a plateau phase approaching an asymptote at 81.2% of applied NAA (equivalent to 0.406 nmol) at 120 h. Within 1 h after application droplets appeared dry on visual inspection. Immediately after droplet drying, 7.0% of the applied NAA was sorbed to the cuticle, but only 0.5% penetrated into the receiver solution, indicating that penetration occurred almost exclusively from the apparently dry deposit. At 120 h, 5.2% of the NAA applied was associated with the deposit and 4.3% with the cuticle. The distribution of maximum rates of penetration was lognormal, but penetration at 120 h followed a normal distribution. Cuticle thickness (estimated 5±25 mm) had no signi®cant effect on NAA penetration. Maximum rates of penetration through pepper fruit and citrus and ®cus leaf cuticles were 4.9-, 2.6-and 0.1-times that through tomato fruit cuticles. At 120 h, penetration averaged 85.5, 79.5 and 34.7% for pepper, citrus and ®cus cuticles, respectively. Extracting epicuticular and embedded waxes increased NAA penetration rates through tomato fruit cuticle more than three-fold, but had little effect on penetration at 120 h (71.0 vs 87.7% for cuticular vs dewaxed cuticular membranes). The maximum penetration rate and total penetration were found to be useful parameters in describing the penetration time-course.