OBTAINING VARIATIONS 99 



tion from the three parent forms, will combine 

 all the hereditary factors of their three special- 

 ized ancestors. 



Among these second-generation hybrids there 

 wiU be found, in all probability — if large num- 

 bers of specimens are grown — some individuals 

 that will combine in superlative degree the quali- 

 ties of gracefulness of plant of one grandparent 

 with the satisfactory arrangement of flowers of 

 the second grandparent and the clear whiteness 

 of blossoms of the third grandparent. 



It is, then, an obvious procedure to save the 

 seed of this individual, and while we must expect 

 wide variation among the plants gro^vn from 

 that seed, there will almost certainly be some 

 among them that will reproduce the combined 

 good qualities of the parent, and further selec- 

 tion along precisely the same line — called "line 

 breeding" — will result in fixing of the type, so 

 that we shall have the variety, hitherto existing 

 only in our imagination, which we have all along 

 been seeking to produce. 



Moreover, not alone shall we have produced 

 a type which combines all the best qualities of 

 the different members of the original balloon- 

 flowers, but this new race wiU almost certainly 

 present these characters in markedly accentu- 

 ated form. The perfected balloon-flower will 



