Lymphoscintigraphy was performed pre- or postoperatively in 42 patients with malignant melanoma of the skin. The test consisted of the injection of 1.5-2.0 mCi 99mTc-labelled antimony sulphide colloid, both intracutaneously and subcutaneously, either around the primary tumour or along its edges or into the centre of the excision scar or contralaterally. The draining lymph-node groups were scanned with a gamma-camera 4-6 hours after the injection. Except for one patient, the radioactive tracer had always disappeared from the injection site and had accumulated in th regional lymph-nodes. In 11 of 28 patients with melanoma on the trunk, lymphatic drainage was to more than one regional lymph-node group. In most cases with cutaneous melanoma on the trunk, lymphatic drainage could not be anatomically predicted, but easily and unequivocally identified by lymphoscintigraphy. On the other hand, lymphatic drainage of a melanoma of the head or neck, as well as of distal segments of the limbs, almost always followed anatomical expectations. Using lymphoscintigraphic findings, prophylactic lymphadenectomy could be performed selectively in "high risk" melanomas.