442 GENETICS OF SOMATIC CELLS 



suitable for the application of fluctuation tests, model experiments with artificial 

 mixtures, and combined in vitro-in vivo approaches. Endocrine tumors do not differ 

 from other neoplasms in principle, but they have one very important advantage: 

 with most other tumors nothing is known about the superimposed growth-controlling 

 devices that regulate cell division in the normal tissues of origin, although they must 

 exist. For this reason, it is difficult to assess the dependence or independence of the 

 neoplastic cells in relation to the controlling forces, except in the most vague and un- 

 certain terms. While the full story of the control of growth is probably not known for 

 any single endocrine tissue either, at the present stage of our knowledge it must be 

 regarded as a tremendous advantage that something is known, and cellular multipli- 

 cation can be made to depend on a single, well-characterized hormone in such systems. 

 As pointed out repeatedly by Furth, 412, 413 endocrine tumors are therefore useful 

 models for the study of cell dependence and autonomy in relation to growth-controlling 

 mechanisms. 



Meanwhile, it may be wise to postpone hasty conclusions about the role of variation 

 and selection in the development of hormone independence. Many attractive models 

 can be borrowed from the field of microbial genetics, and there is no doubt that the evolu- 

 tionary mechanism seems intellectually the most plausible at the present time. Just 

 because of this rationale, caution has to be advocated against reasoning by analogy 

 in the absence of any experimental evidence. Some rather preliminary information 

 does indeed suggest that other factors, such as population size, may have a bearing on 

 hormone dependence versus independence at least in special cases. 725 A certain hor- 

 monal stimulus may be required for growth while the number of tumor cells is small, 

 but this may no longer be necessary after the population has reached a certain critical 

 size. It can be speculated that tumor cells might be capable of manufacturing growth- 

 stimulating substances that have to reach a critical concentration and can then replace 

 the requirement for the exogenous hormonal stimulus. Model experiments involving 

 the mixture of metabolically active but reproductively inactive, X-irradiated cells 

 with a small number of viable cells provide support for such a concept, 1051 suggesting 

 that a progressional change may sometimes occur as an automatic consequence of 

 increasing population size. In another, as yet unpublished study, some estrogen- 

 induced and dependent, interstitial-cell tumors of the testis were studied in our laboratory 

 by serial transplantation into hosts with the genotype of origin. Four tumors were strictly 

 dependent on estrogenic stimulation. They grew well upon transfer to estrogen- 

 pellet-bearing males and females and more slowly but regularly in untreated females. 

 They consistently refused to grow when inoculated into untreated males. After they 

 have reached a certain size in estrogen-treated males, the pellet could be removed, 

 however, without influencing the further growth of the tumor. The tumor continued 

 to grow although it did not become hormone independent, because on the next trans- 

 plantation the same dependence was apparent as previously. The fact that the beha- 

 vior of the large population does not breed true strongly indicates that we are dealing 

 with a physiologic mechanism, perhaps involving self-stimulation of growth of the 



