I examined the epithelial surface, by high resolution scanning electron microscopy, after its recovery from the instillation of artificial tears containing chlorobutanol or benzalkonium chloride. Is the mild epithelial surface compromise observed after occasional use of a chlorobutanol-containing artificial tear more substantial after use of a benzalkonium-containing artificial tear? Two drops of a chlorobutanol- or benzalkonium chloride-containing artificial tear were instilled into the right eye of 6 female gray rabbits (2 kg) at 9:00 p.m. and 9:00 a.m. At the same time six control animals received no eyedrops. All animals were euthanized at 3:00 p.m. and the central region of the corneal epithelium quantitatively assessed using a digitizer pad/computer system. There were up to 5% exfoliating cells evident at the ocular surface in treated rabbits but with no difference between the two products. Controls had no cell exfoliation (< 0.5%). The distribution of surface areas of the squamous cells in the treated eyes was shifted to slightly larger values than in the controls after use of the chlorobutanol-containing product but the number of epithelial cell craters/cell was unchanged from that of the controls. Cell surface areas were shifted to significantly smaller values than controls after use of the benzalkonium chloride-containing product and there were much fewer epithelial cell craters/cell. The results reveal differences in the effects of preservative-containing artificial tears on the squamous cells of the corneal epithelium in a clinically relevant situation.