Summary The study is based on a series of 226 malignant melanomas of the skin in 220 cases examined and treated by a single team during the period 1947 to 1960. The general policy has been outlined and the surgical technique described. The origin of the malignant melanoma has been traced from the melanocytes in the basal layer of the epidermis, and the relationship between the malignant tumour and its benign counterpart has been discussed. Some clinical and histological diagnostic pitfalls have been outlined. Three stages of malignant melanoma have been defined, according to the degree of invasion of the dermis and of tumour formation. Stages 1and 2 are superficial melanomas and have a much better prognosis than Stage 3. Ulceration is a bad prognostic sign. Survey of this series has given the opportunity to discuss most of the ˦tiological aspects of malignant melanomas. There is a two-to-one female preponderance, mainly due to the prevalence of the melanomas on the faces of elderly women and on the legs of women under the age of 50. Such tumours are less malignant than the average, so that the prognosis for women as a whole is better than that for men. Melanomas of the foot are relatively common and have a worse than average prognosis, possibly because of trauma. Melanomas of the trunk have a relatively poor prognosis, as they are usually seen in a more advanced state. Five cases had multiple primary melanomas, and in three of these the two tumours were bilaterally symmetrical. The question of hormonal dependence of melanoma is discussed, particularly with reference to the influence of pregnancy on the eruption of malignancy. Two cases of spontaneous regression of malignant melanoma are described. The evidence that trauma causes malignant melanoma is not good, but the evidence that it may make the prognosis worse is strong. The methods of spread of malignant melanoma are described. The pigmented flare or halo in which the tumour appears is all malignant and seems to spread by the communication of malignant potential to neighbouring epidermal melanocytes. The manifestation of this process is less than its real extent and the process may be discontinuous, giving rise to satellites. Satellites may also be produced by the permeation of the lymphatics in the papillary layer of the dermis. They are of bad prognostic significance. When obvious permeation of dermal lymphatics is seen microscopically near a malignant melanoma the prognosis is much worse. The regional lymph nodes to which tumours of different sites may metastasise are described. The problem of lymphatic spread to the skin of a graft donor area is discussed. Lymphatic tumour emboli may give rise to skin metastases on the route of lymphatic drainage, particularly if there is lymphostasis. A particular study has been made of these lymphatic skin metastases, with reference to their development by chain reaction and also with reference to the need for adequate primary excision. The problem of prophylactic lymph node dissection is discussed, and it is concluded that this treatment can be of use, provided it is done after an interval of two to three weeks from excision of the primary. This tends to allow lymphatic tumour emboli to reach the regional nodes. If simultaneous prophylactic lymph node dissection and excision of the primary are carried out, the block excision should be all in one piece, as otherwise lymphatic skin metastases are likely to develop in the intervening part. The therapeutic removal of lymph node metastases is a marginally useful operation. The experience gained in this study has been embodied in a number of recommendations. When the clinical diagnosis is in doubt, excision biopsy is recommended. Incision biopsy and all forms of deliberate trauma are condemned. The margin of excision of a primary tumour will depend on its site, on its stage of development, and on the age of the patient. After an interval of two to three weeks the regional lymph nodes should be removed if the primary is not a superficial melanoma, and in all cases where lymphatic permeation of the dermis has been histologically observed. In this series every single case has been followed up, and the five-year successful results in the ninety-three determinate cases is 51·6 per cent., which is better than in any other published series, even when allowance is made for selection of cases. Malignant melanoma is not a rare tumour (Table I), and its prognosis is better than for most of the common forms of cancer (Table II).