T. M. SONNEBORN 185 



perhaps less than in variety 1. However, I have already repeat- 

 edly proved the usefulness of this serological method of distin- 

 guishing closely related varieties, such as 4 and 8, which are the 

 most difficult to distinguish in the usual way by their mating 

 specificities, and the serologic identification has never failed to 

 be confirmed by later mating tests. Beale (unpublished) has had 

 similar success in serologically distinguishing varieties 1 and 9. 



There can be little doubt that the practical usefulness of the 

 serological method of identification of varieties will increase as 

 our knowledge of the serotypes is extended to more strains and 

 more varieties, thereby providing more complete sets of antisera 

 for use in this task. In principle, it should become possible even- 

 tually to identify the varieties of P. aurelia in this way with as 

 much assurance as the so-called species or serotypes of Salmo- 

 nella are now identified. But no one should underestimate the 

 magnitude of the task, even if all of the necessary antisera were 

 available. Their number would be very great, perhaps about 20 

 for each of the 16 varieties, or a total of over 300. 



Mating Type Inheritance. Two groups of varieties, A and B, 

 are distinguished by their systems of mating type determination 

 and inheritance. Group A includes varieties 1, 3, 5, 9, and 11 and 

 also probably the varieties 7 and 15 which have been little stud- 

 ied. Group B includes varieties 2, 4, 6, and 8 and also probably 

 the little studied varieties 10, 12, and 14. Nothing is known about 

 mating type inheritance in variety 16. ' A Variety 13, as will appear 

 later, cannot yet be assigned to either group A or B. 



The outstanding features of the group B system are ( 1 ) the 

 conspicuous role played by the cytoplasm in the determination 

 of mating type and (2) the consequent rarity of change of mat- 

 ing type after fertilization. A fertilized animal usually gives rise 

 to a clone of the same mating type as the parent. The group A 

 system shows no conspicuous effect of the cytoplasm on the 

 determination of mating type. Consequently there is no signifi- 

 cant correlation between the mating type of parent and offspring 

 arising after sexual processes. Instead of being a clonal trait, as 

 in group B, mating type is a caryonidal trait in group A; that is, 

 each new macronucleus that arises in a fertilized animal is inde- 



