542 Heritable Spiral Torsions. 



ities in this arrangement certainly occurred, such as 

 split leaves, trifoliate upi)er whorls, fusion of two pairs 

 of leaves by a so-called line of tearing extending over 

 the internode, a phenomenon which shows that the leaves 

 in their early stages were fused in a spiral manner instead 

 of in whorls. Besides these, there were eight plants, i. e., 

 about 2%, which had a distinct local torsion at one of 

 the upper nodes of the same development as in the parent 

 plant. Their leaf spiral involved from 3-6 leaves. Only 

 these plants were allowed to flower, and there was no 

 other Dipsaens laeiniatus in the garden. Five of them 

 produced a sufficient quantity of seed. 



The second generation of this culture occupied the 

 vears 1888 and 1889 and consisted of 435 plants which 

 produced stems and were allowed considerably more 

 space than their predecessors, 12 specimens being grown 

 to each square meter. Nevertheless the result was the 

 same as in the foregoing generation. All the plants 

 formed erect and tall but straight stems, from 2-2Yi 

 meters high. Amongst them there were five specimens 

 which had a somewhat larger local twisted part than 

 occurred in the preceding generation, inasmuch as in 

 these cases from 7-12 leaves were combined into an un- 

 interrupted spiral. They arose from three of the five 

 parents. Less pronounced torsions (at the nodes) which 

 affected from 2-3 or rarely from 4-6 leaves, occurred 

 fairly commonly in this year. I counted the percentages 

 of 5, 5, 13, 13, and 28% of such individuals per parent. 



A general advance had therefore taken place, brought 

 about parti}' b}- selection and partly by the ample room 

 given to the plants ; but whilst the isolation of twisted 

 specimens of Dipsaens sylvestris torsiis resulted imme- 

 diately in a proportion of 34% of fine, almost complete 



