TricotyloHS Half Races. 



387 



of 1892 I obtained only a single henii-tricotyl from about 

 20,000 bought seeds. The seeds of this plant gave rise 

 in 1893 to five tricotyls and two hemi-tricotyls among 

 4000 seedlings, that is, about 0.2 7©. Their seeds were 

 harvested separately, and as the expectation of tricotyls 

 was a small one, great quantities of it were sown. I 

 recorded from 800 to 2900 from each, i. e., altogether 

 about 15,000 seedlings, and 

 found the ratio of tricotyls 

 (and hemi-tricotyls) to be 

 from 0.2 to 0.4%. It did not, 

 therefore, seem justified to 

 make a selection among the 

 individual seed-parents. In 

 1894, 12 tricotyls the stems of 

 which had remained ternarv, 

 and twelve specimens with 

 normal decussate stems were 

 planted out. Several beauti- 

 ful fasciations and occasional 

 cases of spiral torsion oc- 

 curred in this culture (p. 369) . 

 The seeds were harvested sep- 

 arately. In the spring of 1895 

 they again produced only 

 from 0.1 to 0.4% of tricotyls. 

 Seventeen specimens from the 

 seed pans with from 0.2 to 0.4% were planted out about 

 a meter apart, but tl^e seeds produced, in 1896, scarcely 

 any tricotyls and only from five seed-parents, the pro- 

 portion being from 0.3 to 0.7%. 



In tlie course of four generations selection had, there- 

 fore, brought about practically no advance. 



Fig-. 82. Dracocef^halum mol- 

 davkiim. A whole plant. 



