298 Non-Isolablc Races. 



tlie occurrence of bolters is considerably diminished; in 

 one experiment for instance from 7 to 4%.^ Other 

 results point in the same direction. 



Heuze^ in his valuable little book on the oil plants,^ 

 says with regard to the rape (Brassica Napus olcifcra), 

 that in the north of France it should not be sown before 

 the middle of July or after the middle of August, for in 

 the latter case the plants will not be strong enough to sur- 

 vive the winter, and in the former too large a proportion 

 will set seed in the first year. The same thing is true of 

 a whole series of other biennial plants Ixjth cultivated 

 and wild ; those which germinate late become biennial ; 

 of those which germinate early a greater portion become 

 annual, the earlier the sowing or the germination took 

 place. 



In these cases we are not concerned with the induction 

 of bolting by night frosts, or by any other stimulus, but 

 with a case of inherited variability. It is true that the 

 beet possesses this variability also, but the general con- 

 ditions in this species are much complicated thereby. 

 That we have to deal with a phenomenon of inheritance 

 is proved by the fact that the annual form can easily 

 be fixed by selection, without, however, attaining a state 

 of absolute purity. Rimpau sowed the seed of bolters,'" 

 and by always selecting seeds ripened in the first vear, 

 he obtained in the fourth generation a race whose seeds 

 when sown on the 31st of March produced annual ])lants 

 only and which in the fifth generation, when sown on 

 the 5th of April, w^as as constant an annual as the normal 



^ W. Rimpau, Das Aufschiessen dcr Rtinkelriihen, Landwirtsch. 

 Jahrbiicher, i88o, p. 192. 



"L. Heuze, Lcs plantes oleagineuses, Bibliotheque du cultivateiir, 

 Paris, 2d ed., p. 16. 



" Loc. cif., p. 197. 



