KLEIN 1958) that cloned lines of tumor cells with distinct physiological or morphological properties frequently differ from one another in observable characteristics of their karyotypes. The obvious question arises: are the distinct physiological or morphological properties of the various cell lines caused by the differences in the karyotypes, or, are the latter only an additional expression of the great variability of tumor cells? Most workers in this field consider the present evidence insufficient to decide the question in either way. The purpose of this paper is to present further data bearing on the same question. As will be seen, the data favor strongly the assumption of a causal relationship between karyotype and phenotype in a tissue culture line of human neoplastic cells, the strain HeLa (GEY et al. 1952).Several factors may be held responsible for the difficulties encountered in a study of the relationship between karyotype and phenotype in tumor cell populations. One of the factors is the great frequency with which the karyotype changes in populations carrying aneuploid chromosome constitutions. In strain HeLa, for example, one out of a hundred mitoses on the average is abnormal (CHU and GILES 1958). Since most of these abnormal mitoses lead to an unequal distribution of the chromosomes to the daughter cells, new karyotypes are formed at a high frequency. Cloned cell populations contain therefore after a short period of growth a number of variant karyotypes. It was found possible to overcome this difficulty by studying the karyotypes of cloned populations within one to two months after the isolation from the single cell.Another difficulty in this study lies in the great variety of karyotypes that arise as a consequence of the large number of chromosomes and their more or less random reassortment during abnormal mitoses. Due to the great variety of possible chromosome combinations, many different karyotypes may be expected to lead to phenotypes indistinguishable in respect to the mutant character studied. This situation excludes, therefore, a priori, the possibility to find a one-to-one relationship between karyotype and phenotype, which would be the most direct evidence of a causal relationship between karyotype and phenotype.It is, however, possible to look for a part-a1 solution of the problem by studying the relationship between karyotype and phenotype in one direction only, that is,