AbstractFifteen malignant melanomas were subjected to chromosomal analysis, in order to determine whether the number of chromosomes in a tumor cell could be correlated to a clinical prognosis. The protocol involved a direct technique which utilized a Colcemid blockade of spindle formation in mitosis to allow study of metaphase chromosomes. The direct method was chosen to reveal chromosome changes in the tumor cell in situ rather than changes in the cultured cell. Two tumors yielded chromosome spreads which could be counted and correlated with a clinical prognosis. Failure to obtain other adequate chromosome spreads were accounted for by the presence of tissue necrosis and the absence of viable tumor cells.