160 ON BIASTREPSIS IN ITS RELATION TO CULTIVATION. 



in pans kept in the greenhouse, the seedlings being transplanted 

 into pots early in April and planted out in the beds at the beginning 

 of May. But there were only thirty-three plants to four square 

 metres, that is about eight to the square metre. At the end of Oc- 

 tober I found that fourteen of the leaf-rosettes showed spiral phyllo- 

 taxis in the centre, seven showed Vs phyllotaxis and twelve were 

 normally decussate; that is, 42 per cent, had spiral and 21 per cent. 

 \/3 phyllotaxis and 36 per cent, were decussate. In May 1896 1 

 confirmed this result and then all but the spiral plants were weeded 

 out. The six most strongly twisted individuals were selected as 

 seed-bearers and were isolated before flowering. 



The seventh generation, 1897-8, was raised in much the same 

 way. The seed of 1896 was sown in the greenhouse on May 5, 1897; 

 the seedlings were potted off, and about the beginning of July 

 they were planted out, seventy in all, in the beds, sixteen plants to 

 the square metre. 



At the end of May 1898 I found that these consisted of — 



Plants with twisted stems 32 = 46 % 



Va phyllotaxis 21 = 30 % 



„ normal decussate 17 = 24 % 



Total 70 



The proportion of stems in which the twist was to the left or to 

 the right respectively was maintained unaltered in this generation, 

 there being usually about an equal number of the two kinds, as 

 shown in the following table: — 



The whole history of the breed may be summarized in a tabular 

 form as follows: — 



