Reappraisal of follow-up data obtained on patients whose uveal melanomas were treated by enucleation before 1955 has led to the conclusion that the operation may have had an adverse effect, accelerating dissemination and a lethal outcome, especially among patients whose tumors were large and/or contained epitheloid cells. The temporal relationship of tumor deaths to enucleation is demonstrable regardless of whether calculations represent "death density" or "hazard" functions. Use of an inferred natural history model suggests that uveal melanomas are slowly growing neoplasms. An average of seven to eight years is required for small tumors to become large and an additional period of four years before metastatic diseases might be anticipated. The eye is exceptionally vulnerable to great fluctuations in tissue pressure that may account for iatrogenic spread of tumor cells during a variety of manipulations before and during enucleation. Research to prevent dissemination by improved management is indicated.