T. M. SONNEBORN 211 



tection of heterozygotes might have been imperfect, so that his 

 figure is to be considered minimal. Knowing the methods em- 

 ployed, I believe the error cannot be very great. Even if it is 

 assumed, as is by no means justified, that only one-fourth of the 

 heterozygotes were identified, their deviation from the expecta- 

 tion on random mating is still significant at the 1% level of 

 confidence. Accepting the data at their face value, it can be cal- 

 culated that at least 78.5% of the homozygotes arose by inbreed- 

 ing. Altogether I consider these field results to constitute strong 

 evidence for the occurrence of considerable inbreeding in nature 

 in variety 9. 



The preponderance of homozygotes could be due to conjuga- 

 tion between relatives or to autogamy or both. Some calculations 

 which seem to bear on this can be made from data given by 

 Pringle ( 1956 ) in connection with the observations mentioned 

 above. Applying the Poisson formula to the proportion of 30-ml. 

 samples which contained no animals of variety 9, the peak den- 

 sities, found in late summer, amounted to only 25 animals per 

 liter or 1 animal per 40 ml. This calculation is surely subject to 

 some error for it assumes unjustifiably that the animals were 

 randomly distributed in the area sampled; but it is a good first 

 approximation, in spite of the occasional finding of local aggre- 

 gations by other collectors. During a considerable part of the 

 year, winter and spring, few or no variety 9 animals could be 

 found. With such low population densities, even at their peak, 

 the chances of animals of complementary mating type meeting 

 and mating would seem to be low except in so far as local aggre- 

 gations occur. Yet the occurrence of some heterozygotes proves, 

 as Pringle points out, that conjugation must occur sometimes be- 

 tween animals of different clones. The low population densities 

 and the excess of homozygotes on the other hand suggest that 

 autogamy probably also takes place. As mentioned earlier, the 

 invariable existence of autogamy or selfing during senility in 

 the laboratory indicates that there has been strong natural selec- 

 tion to maintain them, and this could be true only if the processes 

 do occur in nature. 



The preceding analysis applies only to what goes on in a 



