A prospective randomized study was initiated at our institution in 1972 to determine the efficacy of routine elective lymphadenectomy in localized (stage I) melanoma. Included in the study were 171 patients, 62 of whom had no lymphadenectomy, 55 of whom had delayed lymphadenectomy, and 54 of whom had immediate lymphadenectomy. No significant difference was found among the three treatment groups with respect to survival or metastasis-free survival. Multifactorial analysis indicated that the level of invasion and the thickness of the lesion were the most important prognostic factors, followed by age (60 years or older), site (legs), and tumor type (nodular). A prognostic index based on these variables was highly predictive of metastasis or death. Even when this score was considered, no significant variation was noted among the three treatment groups. More subsequent complications of melanoma, however, occurred in the group with no lymphadenectomy--36 in this group but only 19 in each of the other treatment groups. This finding was not statistically significant but does indicate that a few additional problems may be associated with leaving regional nodes intact. Further studies are needed, and indeed are being conducted, to determine whether elective lymphadenectomy improves survival sufficiently to offset the costs and the complications associated with this approach.