Between September 1985 and September 1987, 101 eyes with rhegmatogenous retinal detachment and severe (with a classification of at least C-3) proliferative vitreoretinopathy but without prior vitrectomy were treated with vitrectomy and randomized to either a mixture of 20% sulfur hexafluoride gas and air or to 1000 centistokes of silicone oil. Between 50% and 60% of eyes that received silicone oil had visual acuity better than or equal to 5/200 compared with 30% to 40% of the eyes that received sulfur hexafluoride gas (P less than .05). Macula attachment was more frequent in eyes that received silicone oil than in those that received sulfur hexafluoride gas (80% vs 60%, P less than .05). Hypotony was more prevalent in eyes with a detached macula (40% to 50% for sulfur hexafluoride gas vs 25% to 30% for silicone oil) when compared with those with attached maculas (less than 5% for either modality). Keratopathy was more prevalent in eyes with detached maculas (about 55% to 60% for either modality) compared with eyes with attached maculas (25% to 30% for sulfur hexafluoride gas vs 10% to 15% for silicone oil). In a companion article, we show that these differences between a gas tamponade and silicone oil are not found for perfluoropropane gas.