In a previous study by the authors, it was determined that irrigation solutions containing an anti-prostaglandin agent had the capability of preventing the occurrence of blood-ocular barrier disruptions, retinal edema and post-operative inflammation following vitrectomy. However, not only prostaglandins but also leukotrienes, noted for their migratory effect on leukocytes, have cell migratory effect on the retinal pigment epithelial cells. Therefore, irrigation solutions containing an anti-prostaglandin agent and a steroid are expected to have the capability of preventing the occurrence of proliferative vitreoretinopathy and post-operative inflammation after vitrectomy. In this experiment, two kinds of steroids were injected into rabbit vitreous cavities in order to evaluate toxicity and to establish the toxic intravitreal dose of steroids on intraocular tissue. After paracentesis, 36 eyes of 18 rabbits were injected with dexamethasone sodium phosphate or betamethasone sodium phosphate, suspended in 0.3 ml of distilled water. Each steroid has the same molecular weight and almost the same clinical effects. The four eyes of two rabbits received 0.3 ml of physiologic saline as a control. All eyes were observed histologically seven days after the injection. It was proved that each steroid had almost the same toxicity for the same concentration and all steroids at dose less than five mg did not show toxicity on intraocular tissue. With all steroids, 10mg produced localized retinal degeneration at the inferior region of the posterior fundus. Twenty mg caused more extensive retinal degeneration. At a dose of 80 mg each steroid caused corneal edema, degeneration of ciliary bodies, fibrin formation in the vitreous body and degeneration of retinal cells throughout the retina. Considering the osmolarity and pH of each steroid solution, it was concluded that less than five mg of each of the steroids did not cause retinotoxicity, whereas more than 10 mg of any steroid did cause retinotoxicity.