144 EDWIN CARLETON MACDOWELL 



yet the extremes were as high as in the uncrossed race. To 

 explain this one may assume that when the extra factor is more 

 potent, it is more easily weakened by its contact with the 

 normal factor, but that in some cases it is not weakened at all. 

 If there be free variations in the factor it is difficult to under- 

 stand the rigidity with which the normal four bristles are held 

 as the. lower limit of variation. There may be flies with only 

 the normal four bristles in almost any generation of selection, 

 yet, no matter how small a fly may be, the four normal bristles 

 are inevitably found. 



There can be no question as to which of the two main hypo- 

 theses is more fully supported by the interpretations that have 

 been made of the facts. The supporting hypotheses required 

 to make the hypothesis of a single varying factor fit the case 

 are so numerous and, in some cases, so unthinkable as to render 

 the main hypothesis of very slight value. 



It may be claimed that, in time, in a hundred instead of fifty 

 generations, selection could have accumulated enough ger- 

 minal variations to bring the germ plasm again in contro of the 

 numbers of extra bristles. But if there are differences in the 

 germ plasm already present or continuously occurring, a long 

 series of generations is not required to prove their existence. 

 It would seem rather a forced conclusion to claim that the 

 changes that might finally appear in a line that had been se- 

 lected without success for a long time, were due to the long 

 selection. It would not be hard to believe that if one waited 

 long enough, without any selection at all, mutations could be 

 found in almost any material. The point at issue is whether the 

 changes in the germ plasm are continuous or spasmodic, whether 

 they are like the continuous breakers on an ocean beach or the 

 storm-caused splashings of an inland pond. It does not take 

 long to find the breakers at the sea shore; one may spend weeks 

 beside a little lake without seeing a single wave dash against the 

 bank. In these experiments germinal differences were found to 

 exist; their presence was found almost immediately. But after 

 selection had separated the different kinds, no further differences 

 were found. 



