ON BIASTREPSIS IN ITS RELATION TO CULTIVATION. 165 



influence upon the percentage of twisted steins. Under proper 

 treatment tlie plants come on so vigorously in the course of the 

 summer that whatever differences may have originally existed 

 gradually disappear, so that they cannot be detected in the winter 

 or in the second summer after the sowing has taken place. 



Dipsacus sylvestris torsus is strictly biennial, in spite of careful 

 observation 1 have been unable to discover any annual individuals. 

 1 obtained seeds of D. sylvestris from various Botanic Gardens and 

 made large sowings, but in no case did an annual form occur. This 

 is the more remarkable since most biennials (such as Beta, Daucus, 

 Oenothera) produce numerous individuals which 'bolt,' from which 

 an annual breed can easily be raised. Under these circumstances the 

 question arises: What would be the effect of sowing the seed, not 

 in the spring, but in summer or in autumn? and I have endeavoured 

 to find an answer to this question by the experiments of which the 

 following is an account. 



Seeds were sown in the summers of 1892 and 1893, at the be- 

 ginning of June, at the end of July and in the middle of August. 

 The seed used belonged to the fourth or fifth generation of my breed 

 which had already attained 34 per cent, of twisted main stems. 

 The chief result was that the plants obtained by summer-sowing 

 regularly developed their shoots in the following year; but, with the 

 exception of a single individual of the June sowing, the stems were 

 not twisted. The details of the various cultures are as follows: — 



Sowing of June 1, 1893. The seed was that of 1892, from a crop 

 which contained 34 per cent, of twisted stems; it was sown in rows 

 on the bed. The seedlings came up regularly; and up to October 

 superfluous plants were weeded out so soon as they began to touch 

 one another. About twenty plants per square metre remained, so 

 that the condition of space was very favourable (see p. 160); there 

 were 179 plants in 9 square metres. In May 1894 all the rosettes 

 shot up without a single exception. As soon as the phyllotaxis of 

 the shoot could be made out, the plants were examined. One stem 

 was found to be characteristically twisted, another had three-leaved 

 whorls, whilst all the others had the normal decussate phyllotaxis 

 and showed no twisting of the stem. 



All the plants, with the exception of the twisted individual, 

 were dug up. This plant clearly showed itself, in the course of the 

 summer, to be less vigorous than the twisted individuals of a normal 

 culture and it flowered later. Its stem was twisted right up to the 

 inflorescence. 



