THE FLOWER 229 



this : If two parents differing in some fixed characteristic 

 be crossed, the entire offspring, in the first generation, will be 

 like the parent possessing the dominant quality. If all the 

 seed of this generation is planted and carefully protected 

 from foreign pollen, its offspring composing the second 

 generation from the parents will vary in the proportion of 

 | dominants (D, D', line 2 of the diagram) to | recessives (R). 

 Planting all the seeds of the second generation and carefully 

 shielding their progeny from foreign pollen, we get from D, 

 line 2, all pure dominants (D, line 3) that is, plants pro- 

 ducing only their own type, and from R, line 2, all pure 

 recessives (R, line 3). But from each of the two sets of dom- 

 inants, D'D' ', line 2, marked " impure " in the diagram, and 

 so called because their seeds may produce both dominants 

 and recessives, we get the same result as in the second gen- 

 eration, namely: pure dominants (D'D', line 3), pure reces- 

 sives (R'R', line 3), and impure dominants (D"D", D"D", line 

 3). If it were possible to distinguish the seeds of these im- 

 pure dominants before germination and plant them only, for 

 no matter how many generations, the result would always be 

 approximately the same, ] pure dominants, \ pure reces- 

 sives, and | impure dominants capable of producing both 

 dominants and recessives in the proportion of 3 : 1. 



259. Practical applications. Four principles of great 

 importance to plant breeders follow from this law in cases to 

 which it applies: (1) the absence of variation in the first 

 generation of hybrids is no sign that it may not occur later; 

 (2) pure recessives always breed true ; hence, if they show 

 the desired character, no further selection is necessary for 

 that character; (3) pure dominants always breed true, but 

 the distinction between pure and impure is usually not 

 apparent in one generation; (4) the descendants of " im- 

 pure " parents cannot be depended upon to come true to 

 either type, but impure dominants may breed recessives, and 

 vice versa, with the presumption, however, of 3:1 in favor 

 of dominants. 



