BISEXUAL REPRODUCTION 223 



young per litter. At the beginning of the inbreeding experi- 

 ment strains of large, vigorous, rapid-growing and fecund 

 animals were isolated from the general stock, and those 

 characters seem to have been maintained without diminu- 

 tion under the continuous selection exercised in choosing as 

 breeders the largest and best nourished individuals from 

 each litter, but no evidence is forthcoming of further pro- 

 gressive genetic changes. 



In hooded rats inbred, but not exclusively in brother- 

 sister matings, for twenty generations, selection has been 

 made successfully for change of the hooded pattern in op- 

 posite directions, to make the race as white as possible in one 

 line, and as dark as possible in another line. (See Tables 27 

 and 28.) Genetic variability decreased somewhat during 

 the first seven or eight generations, which probably sufficed 

 to eliminate most of the genetic variability originally present 

 in the stock as modifying factors. But subsequently the 

 variability as measured by the standard deviation showed 

 little change up to the end of the experiment in generation 

 21 when the selected races died out owing to the prevalence 

 of disease and infertility. The case seems to be best inter- 

 preted as one in which minor genetic changes are continually 

 occurring, so that selection utilizing these may move the 

 racial mode and mean either in a plus or in a minus direction 

 without encountering impassable limits short of an all white 

 or an all black condition. There is a strong parallelism be- 

 tween the variability of the white-spotting pattern in rats 

 and other mammals and the variability of variegated seed- 

 coat in maize and of variegated foliage in a great many 

 species of plants. In both sets of cases an unstable mosaic 

 of alternative characters exists, pigmentation and nonpig- 

 mentation; somatic variation in the relative proportions of 

 the balanced characters is constantly occurring, and germi- 

 nal variation of a similar nature very commonly occurs at 

 the same time as the somatic variation, so tliat selection on 

 the basis of the somatic variation effects germinal change in 

 the race. The variability (or "mutability") in the case of 



