68 INBREEDING AND OUTBBEEDING 



2. If the parental races are pure, F 2 populations are 

 similar, no matter what F 1 individuals produce them, 

 since all variability in the F l generation is the result of 

 varying external conditions. 



3. The variability of the F 2 populations produced 

 from such crosses should be much greater than that of the 

 F l populations, and if a sufficient number of individuals 

 are produced the grandparental types should be recov- 

 ered. The fulfillment of this condition comes about from 

 the general laws of segregation of factors in F 1 and their 

 recombination in F 2 . 



4. In certain cases F 2 individuals should be produced 

 showing a greater or a less extreme development of the 

 character complex than either grandparent. This is 

 merely the result of recombination of modifiers, as was 

 explained above. 



5. Individuals of different types from the F 2 genera- 

 tion should produce populations differing in type. The 

 idea on which this statement is based is, of course, that all 

 F 2 individuals are not alike in their inherited constitution 

 and therefore must breed differently. 



6. Individuals either of the same or of different types 

 chosen from the F 2 generation should give F 3 populations 

 differing in the amount of their variability. This con- 

 clusion depends on the fact that some individuals in the 

 F 2 generation will be heterozygous for many factors and 

 some heterozygous for only a few factors. 



Such are the conditions which must be fulfilled by 

 crosses exhibiting size differences if we are to visualize 

 their inheritance in the same way as we visualize the in- 

 heritance of qualitative characters such as color. If the 

 size differences are controlled by numerous germ-cell f ac- 



