224 GENETICS AND EUGENICS 



plants with variegated seed-coat or foliage extends into end 

 stages of the series which are wholly colored or wholly color- 

 less, which stages seem to be more stable than the inter- 

 mediate (mosaic) stages, as pointed out by Emerson. It is 

 to be regretted that in the selection experiments with rats 

 similar end stages were not reached before the selected races 

 perished. In the case of Dutch rabbits (Figs. 138-140) the all- 

 white condition has been recorded once, and the all-colored 

 condition is often found in animals known to be either 

 heterozygous or homozygous for some form of white spotting. 



Two different explanations of cases of this class in animals 

 and plants have been suggested. (1) On one view the chief 

 genetic locus mutates frequently producing multiple alle- 

 lomorphs more or less stable (Emerson), but these multiple 

 allelomorphs may be supplemented in action by minor modi- 

 fying genes (Hayes). (2) On another view the chief gene is 

 as stable as other genes and the ordinary genetic variability 

 is due exclusively to modifying genes (MacDowell, Pearl, 

 Sturtevant) . If the chief gene is really less stable in the case 

 of these mosaic characters than in ordinary cases, as the 

 descriptive term used by DeVries, "ever sporting char- 

 acters," would seem to imply, at least in the case of plants, 

 it may be because a mosaic condition exists at the genetic 

 locus itself. In variegated plants the character of the mosaic 

 in particular parts of the plant corresponds roughly with 

 the character of variegation transmitted by flowers arising 

 in those same parts of the plant whether egg-cells or pollen- 

 cells are the vehicles of transmission, which suggests actual 

 variation in the genetic locus involved rather than change in 

 modifying genes. (See Figs. 135, 136.) 



MacDowell inbred, brother with sister, a race of Droso- 

 phila possessing a recessive Mendelian character, extra 

 bristles, for 49 generations, selecting meanwhile in different 

 lines for high and for low number of extra bristles. For about 

 eight generations the selection was effective after which no 

 material change was observed attributable to the selection. 

 MacDowell concludes that at the beginning of the experi- 



