536 Heritable Spiral Torsions. 
importance, inasmuch as it determines the length of the 
life of the plant up to the moment of the production of 
the stem. The longer this period continues under favor- 
able circumstances the greater is the likelihood of the 
leaves becoming spirally arranged. 
Sowings, made in summer or early autumn, which 
gave rise to stems in the next year, reduced the prospect 
of obtaining torsions almost to nil. On the other hand, 
autumn sowings which do not give rise to stems until 
the summer after the next, contain very large numbers 
of twisted specimens; so that if the seeds are sown late 
in the autumn the proportion of plants with a spiral ar- 
rangement of their leaves is even greater than amongst 
plants raised from seeds sown in ths spring. 
Little effect is produced upon the result of a culture 
by sowing in March or in April or even in the begin- 
ning of May. Also it does not matter much whether 
the sowing is carried out in the greenhouse in pans and 
the seedlings planted out later into the beds, or whether 
the seeds are sown where they are to grow. For various 
reasons I have for many years preferred the former 
method, as it is more convenient and safer, especially in 
dry springs. 
A good loose soil with a strong manure rich in nitrates 
seems to be an essential condition. On unmanured sandy 
soil even the best seeds do not produce twisted individuals, 
and on hard or barren soil the proportion is considerably 
diminished. 
It is possible to confine the cycle of life of Dipsacus 
sylvestris torsus within the limits of a year by sowing 
the seeds under favorable circumstances immediately 
after they are ripe. In this way a generation can be 
grown every year, and an annual twisted race might pos- 
