296 Non-Isolablc Races. 
regions as pure biennials, in others as annuals, and in 
still others in a mixture of these two forms. 1 
Inasmuch therefore as the biennial habit is to be re- 
garded as the character of the species and the annual 
habit as the anomaly, the latter is likely to follow the 
general rule according to which the development of the 
anomaly is favored by improved conditions of life. And 
the experiments which I propose to describe in this sec- 
tion prove the correctness of this view. 
However, there is an apparent contradiction, for, as 
is well known, RIMPAU has shown in the case of the 
beet that every retardation or interruption of the growth, 
whether it occurs during germination or just after the 
seed comes up or at a later stage of the development of 
the plant, favors the production of the seed in the first 
year of the plant's life. 2 
But in this case it only appears that we are dealing 
with conditions favorable to the production of the anom- 
aly whereas in reality we are concerned with the stimulus 
necessary for the manifestation of this bolting. As it 
is not very easy to make this difference clear I shall select 
an instance of a pure biennial race 3 which lacks the power 
of giving rise to annual specimens. I refer to my cul- 
tures of Dipsacus sylvestris. This race can be sown at 
any period of the year, and the plants will always remain 
rosettes until the end of the next winter and develop a 
stem in the spring of their second year. According to 
whether the sowing was made in the spring or in the 
summer or not till autumn are the rosettes vigorous or 
1 Instances of this are given by J COSTANTIN, Les vcgctanx ct 
Ics milieux cosmiques, Paris, 1898, pp. 28 f. 
2 Landw. Jahrbilcher, passim, 1880, p. 194. 
3 On Biastrepsis and Its Relation to Cultivation^ Annals of Bot- 
any, Vol. XIII, No. LI, Sept. 1899, p. 395. 
