T. M. SONNEBORN 213 



immature period in order to obtain a satisfactory picture of the 

 breeding relations. 



The diversity of conditions within variety 1 may here be noted 

 also for certain other features. In one stream in Woodstock, 

 Maryland, I found many years ago a strain pure for mating type 

 I. No such strain has been found among the many collections 

 of variety 1 from the rest of the world, including Asia, Aus- 

 tralia, South America, and Europe. Its significance for the breed- 

 ing system is the same as set forth above for the pure strain of 

 type XIII in variety 7. Of special interest, however, are two 

 further pieces of information. First, I collected from this same 

 source several times over a period of years and always found 

 there this same strain and no other variety 1 (although variety 

 2 was also always found with it). This shows that the strain can 

 and does persist for years in nature by repeated autogamous 

 inbreeding, a conclusion supported by 18 years of maintenance 

 in the laboratory. The same facts also indicate how constant a 

 local population can remain for a long period. Second, Beale has 

 examined the serotype loci of representatives of this population 

 and finds them to be entirely homozygous at all three major sero- 

 type loci. This is in marked contrast to his observations of hetero- 

 zygosity at these loci in representatives of other populations that 

 contain both mating types. It is therefore clear that within variety 

 1 there are populations that inbreed closely and almost exclu- 

 sively by autogamy and other populations that at least occa- 

 sionally undergo conjugation. The constancy of the Wood- 

 stock population further suggests that the conjugations in other 

 local populations are probably as a rule within the population. 

 Migrations and interbreeding between populations seem to be 

 relatively rare. 



There are differences, as well as similarities, in the life features 

 of varieties 1, 2, and 3. Variety 2 belongs to group B, whereas 

 varieties 1 and 3 belong to group A. Variety 2, like variety 9, 

 seems to show relatively little F2 mortality in the strain crosses 

 that have been examined. On the other hand, variety 1 shows 

 exceedingly high F2 autogamous mortality after crossing certain 



