ON BIASTREPSIS IN ITS RELATION TO CULTIVATION. 161 



Whilst at first a large number of plants, crowded together, were 

 used with imperfect success, in the later years fewer plants with 

 plenty of space have afforded much more satisfactory results. This 

 improvement is due in part to the more favourable cultural methods, 

 in part to the continuous selection; it is impossible in this case, as 

 usually in other such cases, to discriminate between the effects due 

 to these two causes respectively. It will, however, be shown in the 

 next section that the better cultural methods were of considerable 

 importance in bringing about the result. 



B. The influence of space. 



The condition most essential to the successful cultivation of 

 these twisted plants is that each plant shall have sufficient room 

 in which to develop freely; the plants must neither touch nor over- 

 shadow each other. This result is clearly indicated by the various 

 cultures previously described. 



It is instructive, in connexion with this point, to compare the 

 plants growing on the borders of a bed with those growing in the 

 middle. Whenever the space becomes insufficient, the twisted stems 

 are mostly or even entirely confined to the plants growing on the 

 borders: this was the case in the second and third generations, when 

 there were about fifty plants to the square metre. On the other 

 hand, the occurrence of spiral rosettes in the plants on the borders 

 and the absence of them from those in the middle of the bed, is one 

 of the best indications whether or not the plants have sufficient 

 space allotted to them. 



In order to demonstrate in a simple manner the truth of the 

 above statement, I instituted the following experiments. 



In 1889 seeds, obtained from the second generation in 1887, were 

 sown on two neighbouring beds and in the same manner. When, in 

 June, the plants began to touch each other, they were not thinned 

 out to an equal extent; on the one bed 300 plants were left, on the 

 other 540. As each bed had an area of twelve square metres, there 

 were in the one case twenty-five plants to the square metre, in the 

 other forty-five. On examining the elongated shoots in May of the 

 following year the results, which were widely different in the two 

 cases, were — 



Bed, 25 plants to square metre 6 % twisted stems. 



5 % phyllotaxis Vs- 

 Bed, 45 plants to square metre 1 % twisted stems. 



1 % phyllotaxis Vb- 

 II 



