The biological effects of four commercially available artificial tear formulations were evaluated using sensitive in vitro techniques. Two of the formulations contained ingredients implicated in cell damage; the other two products were not chemically preserved, and their components have not been reported to damage corneal tissue. We assayed the effects of these formulations on viability, morphology, and physiology in corneal cell (SIRC) cultures. Their effect on the hydration of excised rabbit corneas was also determined. In all formulations, cell viability declined with time relative to control cells, but the time course varied significantly. Viability remained at 100% for 6 h in an unpreserved carboxymethylcellulosebased product (CMC-U), and decreased to 50% after >16 hours. Viability decreased to 50% in 3 h for the other unpreserved, polyvinyl alcohol-based product (PVA-U), and in 1 h for a hydroxypropylmethylcellulose formulation (HPMC-P) that contains edetate disodium (EDTA). Cells in a preserved formulation (PVA-P), using polyvinyl alcohol as the polymer and containing EDTA and benzalkonium chloride (BAK), failed to survive even 15 min of treatment. Overall, cells treated with the unpreserved products were nearly indistinguishable from those in the control solution with respect to morphology, electrophysiology, and corneal hydration. Also, the relative ranking from least to most deleterious (control < CMC-U < PVA-U < HPMC-P < PVA-P) was consistent across several measures. Of the preserved formulations, HPMC-P, which contains the chelating agent EDTA as an additive, was less damaging than was PVA-P, which contains two chemicals, EDTA and BAK, that reportedly damage cells.