The best method of ascertaining the properties that characterize a malignant tissue would be to transform in vitro a strain of cells of a known type into cells capable of producing sarcoma or carcinoma, and to study the changes undergone by the strain. The comparison of the normal and the malignant strains would show what characteristics are specific of the state of malignancy. As we possess pure cultures of normal cells of several types, and as the contaminating agent of Rous sarcoma renders feasible the transformation in vivo of a normal cell into a malignant one, an attempt could be made to change in vitro normal cells into sarcoma cells. The first experiments consisted in determining the nature of the cells that are susceptible to the agent of Rous sarcoma. It was obvious that the strain selected for the attempt at contamination must be of the same type as the