The aim of this study was to evaluate the sentinel node biopsy (SNB) technique for melanoma using both radiocolloid and blue dye in 241 clinically N0 patients with melanomas >1.0 mm, or thinner lesions exhibiting regression/ulceration. We showed that an increase in injected radioactivity increased both the number of visualized nodes at lymphoscintigraphy and the number of SNs removed surgically. At least one SN was removed in 98% (236) of patients, and all nodes were identified with the probe. Seventy-four per cent of the 194 patients injected with blue dye had stained SNs. In 46% (144) of the lymph node basins, there was a discrepancy between the nodes visualized at lymphoscintigraphy and the nodes removed at surgery. There were 38 unusually located nodes. Only eight of these were removed surgically; none contained metastases. SN metastases were detected in 22% (53) of patients. There were nine haematoxylin and eosin (HE)-negatives, all of which were found by immunohistochemistry. The false negative rate for the SNB procedure was 4% (2/55). The complication rate was 6% after SNB and 29% after complete node dissection. In conclusion, SN status is a strong prognostic factor in melanoma patients, and SNB has made the approach to radical lymphadenectomy more rational.