The results of the studies on regression reviewed here are not comparable for two reasons. The definitions of regression as well as the definitions of an unfavorable course differed considerably. Before a definite word can be said about the prognostic implications of histological features of regression in melanoma, a larger study correlating regression with overall survival of patients with melanoma has to be undertaken. Only melanomas with unequivocal regression identified by reliable and repeatable criteria should be included. Similarly, research on pathophysiology of regression requires precise and repeatable criteria for recognition of regression before meaningful results can be expected.