T. M. SONNEBORN 233 



known. Moreover, observations snch as those I made over a 

 period of years on the population of the Woodstock stream, to- 

 gether with Beale's observations on the occurrence of only one 

 allele at each serotype locus in the samples taken in various years, 

 show how rare must be the effective migration of a foreign strain 

 into a preexisting population. In this case, moreover, the observa- 

 tions were made on a population of variety 1 in the midst of a 

 region well within the range of the variety and where many other 

 populations of the same variety were known to occur. If migra- 

 tions from one body of water to a neighboring one are so rare, 

 then migrations between continents must be exceedingly rare or 

 nonexistent except on a geologic time scale. The fact that strains 

 of the same variety (such as 1, 2, 4, and 6) from different con- 

 tinents can still interbreed freely when brought together in the 

 laboratory implies that the widely distributed varieties must be 

 evolving very slowly in spite of long isolation. Nevertheless, with 

 each local population virtually locked up in the tight compart- 

 ment of a river system or body of still water from which leakage 

 is slow and erratic, the stage would seem to be set for abundant 

 evolutionary divergence. 



Differences in mating type specificity and differing conditions 

 for mating reactivity (temperature, light), once they arise, con- 

 stitute such effective barriers to interbreeding that different vari- 

 eties can and do intimately coexist in the same body of water 

 without losing their integrity. Even when the mating type speci- 

 ficities are only slightly different and interbreeding is possible, 

 as in the case of varieties 4 and 8, the varieties are found to co- 

 exist in the same body of water in nature. From these observa- 

 tions, it would seem that even relatively slight changes in mating 

 type specificity could lead to isolation of a new variety. Probably 

 the inbreeding propensities discussed earlier assist such isolation. 

 Outbreeders might not be so easily isolated in this way. Another 

 related possible isolating factor is assortative mating with respect 

 to body size (Jennings, 1911). There is some evidence that body 

 size is in a general way correlated inversely with the rate at 

 which the processes of meiosis and fertilization occur and marked 



