The Differentiation of Populations \ 235 



the western one by a 25-mile gap and from the eastern one by an 8- 

 mile gap. 



The frequency of blue-flowered individuals in samples from these 

 "variable areas" is shown in Fig. 10.12. This frequency distribution 

 of phenotypes is reminiscent of the theoretical gene-frequency dis- 

 tributions in small populations subjected to some selection pressure 

 (see Fig. 6.9). The curve in Fig. 10.12 cannot be fully interpreted 

 because the genetic bases of the observed variation are unknown. 

 However, the U shape is strongly suggestive of populations suffi- 

 ciently small for considerable random fi.xation and loss of genes 

 producing blue flowers to have occurred. It seems likely that selec- 

 tion, drift, and isolation by distance all interact to produce the ob- 

 served pattern of variation in Linanthiis. Some unknown selective 

 factor probably gives a small advantage to blue-flowered plants in 

 the "variable areas," while the small effective population size (cal- 

 culated by Wright to be 14 to 27 plants ) permits considerable drift. 

 In addition, the usual desert year (much drier than 1941) would 



Fig. 1 0.1 1 I Map showing composition, with respect to flower color, of 

 samples of Linantlius parrijae from an area in southern California. Black 

 sectors, blue flowers; white sectors, white flowers. ( From Epling and 

 Dobzhansky, 1942, Genetics 27, and after Ahrams, 1951, Illustrated Flora 

 of the Pacific States, Stanford University Press. ) 



