4 1 8 De Vries. — On Diastrepsis in its 
shoots which showed no trace of twisting or of three-leaved 
whorls. At Nancy, as at Amsterdam, the production of seed 
within a year could be induced, but at the expense of bia- 
strepsis. 
The seed, gathered at Nancy, sown in 1897, gave in the 
following summer only seven twisted stems, and ten with 
three-leaved whorls, out of about ico shoots. This remarkable 
state of affairs must probably be ascribed, as Professor Le 
Monnier suggests, to the pollination of the twisted seed- 
bearing plants of 1896 by pollen brought by insects from 
other plants of Dipsacus growing at a distance. This experi- 
ence is the more important since I have myself observed that 
an interval of 100 metres is often insufficient to prevent the 
crossing of two varieties of the same species if in flower at 
the same time. 
Conclusions. 
1. The seed of Dipsacus sylvestris torsus yields, under 
proper cultivation, a progeny of which about one-third have 
twisted stems. This proportion was first attained in the 
fourth generation, and since then it has rather increased than 
diminished on the whole (see p. 404). 
In addition, there occur plants with three-leaved whorls, 
with divided leaves, or with local twisting of the lateral branches, 
and occasionally the other anomalies which I have described 
in my monograph 1 . 
2. The development of biastrepsis, that is, the transition 
from decussate to spiral phyllotaxis, depends not only upon 
the hereditary properties of the individual latent in the seed, 
but also in a high degree upon the external conditions under 
which the individual develops. 
3. The more favourable the conditions of life, and con- 
1 In addition to the references given on pp. 396 and 397, I may mention the 
following papers which I have published on this subject : Ber. d. deutsch. bot. Ges., 
xii, 1894 ; and Kruidkundig Jaarboek van het Genootschap Dodonaea in Gent, 
iii, 1891, and iv, 1892. 
