It has been believed that 1% rose bengal does not stain normal, healthy cells but rather stains degenerated or dead cells and mucous strands. In contrast to this conventional knowledge, we discovered that both commercial additive-containing and additive-free rose bengal solutions stained four different types of healthy cultured cells, including rabbit corneal epithelial cells. Rose bengal staining was rapid, dose dependent, predominantly nuclear, and detectable with the naked eye at concentrations as low as 0.05% and 0.025% for the commercial additive-containing or additive-free solutions, respectively, and with the fluorescence microscope at a concentration of 0.001%. It is surprising to discover that rose bengal is not a vital dye; after staining, cells actually lost vitality, as evidenced by instant morphologic changes, subsequent loss of cellular motility, cell detachment, and cell death. Such an intrinsic toxic effect was augmented by light exposure. The rose bengal staining of live as well as detergent-treated (Triton X-100) cells could be blocked by such tear components as mucin and albumin, suggesting that normally negative rose bengal staining is due to the protective function of the preocular tear film, ie, staining is not dictated by lack of cell vitality. These data indicate that rose bengal staining ensues whenever there is poor protection of surface epithelium by the preocular tear film; this represents a new interpretation for rose bengal stains seen in various ocular surface disorders.