T. M. SONNEBORN 253 



further analysis he called for. In my opinion, his suggestion is 

 unsatisfactory, although the genotype probably does play some 

 part in the system. The chief difficulties with Kimball's proposal 

 are that it provides no insight into the basis for the choice be- 

 tween the two alternative types in each synclone or into the 

 mechanism by which either of the alternatives can be inherited 

 through great numbers of fissions. 



These difficulties are precisely the ones which are accounted for 

 by the analysis of the group B method of mating type inheritance 

 in P. aurelia. In extending this to variety 1 of P. bnrsaria, four 

 instead of two cytoplasmic differentiators are assumed in order 

 to meet the requirements of four instead of two mating types. In 

 a steady state system, these four would stand in a hierarchy of 

 relative concentrations in each mating type, and the one initially 

 in highest concentration would be the one corresponding to the 

 expressed mating type. However, the relative concentrations of 

 the other three might well vary in different synclones of the same 

 mating type. With such a system of cytoplasmic constitutions, 

 the cytoplasmic differentiator present in second highest concen- 

 tration could be the one which determines the alternative mating 

 type producible by self-differentiation. On this view, there are in 

 each clone very rare occasions on which the two most concen- 

 trated cytoplasmic differentiators exchange relative positions in 

 the hierarchy. Then, when age renders the macronuclei suscep- 

 tible to redifferentiation, those animals which have the changed 

 hierarchy of differentiators change mating type. As in P. aurelia, 

 the determination of the specific hierarchy in any synclone or 

 subclone is a resultant of the action of the genotype and of ex- 

 ternal conditions. 



This is perhaps not the appropriate place to develop in detail 

 how the previously unexplained peculiar features of mating type 

 inheritance in P. bursaria become intelligible in the light of this 

 approach to the problems; but, to indicate the fruitfulness of the 

 hypothesis, one further example will be cited. Jennings ( 1942, 

 Table 2) lists five crosses between mating types B and C, with 

 different clones involved in each cross. One of the five crosses 

 stands out as giving unusual results. Instead of the expected re- 



