444 GENETICS OF SOMATIC CELLS 



exudate environment. Experiments carried out with the specific purpose of studying 

 this question 713 supported the selection of preexistent variants. The main point was 

 that convertible or inconvertible lines could be produced from the same original 

 tumor at will, simply by modifying the size of the first inoculum in the beginning of the 

 conversion experiment. When very small inocula were used, convertible lines could 

 be changed into inconvertible. This can only be explained by assuming that the variants 

 were already present in the original population of the solid tumor in a small proportion. 

 Extrapolation of model experiments with artificial mixtures led to a tentative estimate 

 of a probable frequency of the order of 4 x 10 ~ 7 . 



These studies may be considered together with the earlier work of Coman, 218 

 who found that tumor cells were characterized by decreased cellular adhesiveness, and 

 the more recent results of Ambrose et al., 17 who showed that their surface carried an 

 increased negative electrical charge when compared to homologous normal cells. 

 Also relevant are the results of Abercrombie and Heaysman, 1 who observed that surface 

 movements of normal fibroblasts in culture will be inhibited upon contact with other 

 fibroblasts, leading to adherence of the cells, while such contact inhibition was largely 

 or completely absent in cultures of sarcoma cells. Taken together, all these data point 

 toward a relationship between changed surface characteristics of neoplastic cells and 

 their tendency to invade and metastasize. Also suggesting a surface change are the 

 findings of Wolff and his school, 1399 who showed that dissociated embryonic cells of 

 different species associate specifically in vitro and build up real organ chimeras that 

 contain a mixture of cells derived from the homologous tissues of the different animal 

 species used. On the other hand, malignant cells appeared to be devoid of such a 

 surface-recognition factor since they invaded all kinds of embryonic tissues without 

 discrimination when tested in a similar system in vitro. The fact that it was possible, 

 in the experiments on ascites conversion discussed in the previous paragraph, to select 

 tumor sublines that differed from each other with regard to their surface characteristics 

 as well as their ability to invade, to metastasize, and to proliferate in the ascitic form, 

 indicates that the surface characteristic in question is not an all-or-none phenomenon 

 but may be expressed to different degrees and is liable to undergo progression in estab- 

 lished tumors through variation and selection. It is to the merit of Foulds 399, 401 

 to have emphasized clearly that progression can proceed independently with regard to 

 different cellular characteristics. Changes in surface and invasive properties do not 

 necessarily have to be correlated with changes in other qualities of the neoplastic cells, 

 such as rate of growth, hormonal dependence, antigenic behavior, and the like. 

 On the other hand, recent work on tumor viruses in vitro, particularly the associations 

 between Rous virus and chicken fibroblast 1080 or iris epithelium 326 and the system of 

 polyoma virus and fibroblast in hamster and mouse 1350 has demonstrated that a tumor 

 virus may impose the entire neoplastic phenotype, including the ability to grow as a 

 tumor upon reimplantation into an appropriate animal host, upon the recipient 

 normal cell within a comparatively short time. Simultaneously with this change, the 

 surface properties and the mutual relationships of the cells are altered, as indicated 



