We analyzed uveal melanoma metastases in a group of 41 patients who received 20 Gy of preenucleation radiation in a Northern California Oncology Group preliminary phase I/II study, and compared their survival rates with a retrospective control group of 31 patients with characteristics matching the entrance criteria but treated with enucleation alone. Using the Cox proportional hazards model, we found that increased tumor diameter, mixed or epithelioid cell type, and radiation adversely affected survival. In vivo studies of cell cycling indicated that 20 Gy of preenucleation radiation appeared to diminish the reproductive integrity of the tumor cells. It is most likely that the failure of preenucleation irradiation to prolong patient survival was because of micrometastases that occurred before treatment.