350 CELL HEREDITY 



and all 1"' , s whose cvtoplasm came from the D parent, were a mixture 

 ot 601) and 90D. I'hus, the new allele which entered at conjugation 

 was expressed in a few fissions, hut whether the g gene or the d gene was 

 chosen depended on the prior state of the cytoplasm. After many divi- 

 sions, those F, animals maintained at 25° C' all developed the G antigens; 

 and genetically identical animals kept at 29° C all developed D antigens. 

 This beautiful experiment demonstrates dramatically the specific contri- 

 butions of gene and cytoplasm in determination of antigenic type. 



Now let us compare these results in a formal way (ignoring the many 

 biological differences between the systems) with phase variation in 

 Salmonella. In both systems, more antigenic potentialities exist in the 

 organism than are expressed at any one time, with each potentiality 

 determined by a different gene. Consequently, in both systems a 

 mechanism is necessary for choosing among alternatives. The Parame- 

 cium analysis demonstrates that this choice operates via the cytoplasm, 

 but reveals nothing about the origin of the cytoplasmic differences. The 

 Salmonella analysis demonstrates the involvement of a chromosomal com- 

 ponent in the choice between expression of the //, and Hy genes, but 

 presumably requires that the effect of this element at the //j locus be 

 mediated via the cytoplasm for it to reach the H^ locus. Conceivably, 

 the two methods of analysis have uncovered two aspects of the same 

 mechanism, there being both a genetic and an environmentally deter- 

 mined component involved in both systems. This suggestion is related 

 to the speculations (p. 331) about the bipartite nature of repressors, 

 one part being of direct genetic origin, the other part a metabolite. If 

 the rate-limiting factor in antigenic change is the metabolite component, 

 as may be likely in the temperature system in Paramecium, the existence 

 of a genetic controlling element could not be detected. 



Much of the discussion in the literature has centered on the possible 

 nature of the cytoplasmic component elicited by the environment, which 

 seems to have the dual role of inducing the formation of one protein and 

 repressing formation of all the others. There has been no evidence 

 whether the choice occurs at the chromosome level or at the level of 

 protein formation, but the prevalent view has been the latter. On the 

 controlling-element hypothesis, these are not mutually exclusive alterna- 

 tives. A controlling element might activate one locus and inactivate an- 

 other with respect to production of a gene product (e.g., part of a 

 repressor) which itself acts at the level of protein synthesis. In an 

 "ordered" system like antigen determination in Paramecium, (in contrast 

 to the "disordered' systems of maize and Salmonella involving high 

 spontaneous mutability), an additional concept is required, namely, a 

 feedback from the cytoplasm to activate the controlling element. 



It must be evident to the reader how remote many of the biological 



