In our experimental study on 53 rabbits we compared the amount of corneal endothelial damage caused by radial keratotomy (RK) referred to (1) the number of incisions (4, 8, or 16), (2) the postoperative interval (0 h, 48 h) and (3) the direction of the incision [centripetal (cp), centrifugal (cf)]. The endothelial damage was quantified by means of the Janus green photometry technique. Morphological changes were evaluated by scanning electron microscopy (SEM). Depending on the group examined, we found endothelial damage extending over 3-7% of an analysed surface of 64 mm2. One perforation caused endothelial damage of up to 17% of the surface examined. Increasing the number of incisions from 4 to 8 or 16 resulted in a statistically significant increase in the amount of endothelial damage (4.2%, 5.1%, 5.8%; P less than 0.05). At 0 h it was significantly higher than after 48 h (5.5%, 4.6%; P less than 0.05). The direction of the incision had no statistically significant influence in our study (zp: 5.2%, zf: 4.9%). The morphological changes in the rabbit corneal endothelium examined directly after the RK procedure were ruptures in the cell membranes, loss of cells, and posterior corneal protrusions beneath the incisions. After 48 h, we found fewer damaged cells and no denuded Descemet's membranes, but larger polymorphy of the cells and a numerical increase in the microvilli of the cells surrounding the damaged cells. Our results support the crucial argument against RK: the alteration and destabilization of healthy corneal tissue up to the endothelium.