T. M. SONNEBORN 227 



killer trait and its various mutant forms, cannot be maintained in 

 the animals. The differences between groups A and B with re- 

 spect to the serotypes and mating types have been commented 

 upon by several authors (Beale, 1954; Nanney, 1954, 1956b; Son- 

 neborn, 1947, 1954b ) . Here I shall discuss only the mating types, 

 for they are of special importance in relation to breeding systems 

 and evolution. 



The group A and group B systems of mating type determina- 

 tion and inheritance have been characterized in detail from vari- 

 ous points of view at several places in the present paper. On the 

 basis of these facts an attempt will be made here only to formu- 

 late these systems in an abstract and general way in order to see 

 how many and what kinds of changes would be required to evolve 

 one from the other. In making this formulation, I shall pass rap- 

 idly over the points on which evidence is good and interpretation 

 is fairly straightforward, but shall say more about certain other 

 points. 



On a few points, common to both systems, there can be little if 

 any doubt. These are: ( 1) the mating types are controlled by the 

 macronucleus; (2) except in certain mutants, all macronuclei 

 carry a genotype for both mating types; (3) usually a macronu- 

 cleus becomes differentiated early in its development from a prod- 

 uct of the syncaryon so that only one of its mating type poten- 

 tialities will thereafter come to expression; (4) this macronuclear 

 differentiation is permanently or long irreversible and, as a result, 

 a caryonide is commonly pure for one mating type. Nanney 

 (1956b) interprets the macronuclear differentiation as an intra- 

 nuclear steady state between products of the alternative genes 

 for different mating types. 



The differences between the group A and B systems involve 

 the agents that bring about macronuclear differentiation and the 

 genetic control of these agents. In group B, the cytoplasm clearly 

 is the agent of macronuclear determination, and the cytoplasmic 

 agent is clearly controlled by that macronuclear mating type geno- 

 type which is being expressed. This is what results in the usual 

 correlation in mating type between the two caryonides of a clone 

 and between both of them and the mating type of the sexual 



