276 BOTANICAL GAZETTE [octobee 



diminutions of the laeta type may, at least in part, be considered 

 as influenced by neighboring, characters. The combination 

 0. cana nanella has occurred in many instances during these experi- 

 ments, either from the crosses of O. cana with the dwarfs of my 

 race, or as occasional mutations in other cases. A repetition of my 

 experiments, in which O. cana nanella would be chosen for ferti- 

 Hzation by the other species or mutants used, would probably give 

 the material to decide these questions. 



Putting aside all of these more special considerations, we may 

 conclude that 0. cana behaves in its crosses, as well as after self- 

 fertilization, exactly after the manner of 0. scintillans and 0. lata. 

 From this fact and from the cross LamarckianaX.pallescens, pre- 

 viously mentioned, it seems probable that the other new dimorphic 

 mutants, namely 0. pallescens, 0. liquida, and 0. Lactuca, will 

 follow the same rule if tried in the same way. This rule is evi- 

 dently independent of the external types of their characters, but 

 must depend upon internal properties of their hereditary qualities.''' 



I will now give briefly the necessary details concerning the 

 crosses mentioned in tables IX and X. As already mentioned, 

 all these crosses were made in 1913 with pjants of the same origin. 

 For every cross a single specimen was chosen and some flowers on 

 the lower part of its main spike were castrated. The seeds were 

 sown in February 1914, the seedlings transplanted into boxes, and 

 from these, as a rule, about 25 specimens of each culture were 

 placed in a bed in April and May, giving them a good soil and 

 light exposure and plenty of space to insure a vigorous develop- 

 ment until the time of flowering and of fruiting. 



0. canaXO. biennis Chicago. — A group of 71 specimens, all 

 of which produced high stems and 25 of which have flowered. 

 One was a mutant, combining the gray narrow leaves of 0. cana 

 with the marks of the stature, foliage, and flowers of O. lata. 

 The plants of the type of 0. cana were exactly like pure 0. cana 

 at the beginning of the flowering period, in July, when they had 

 reached a height of 70 cm. The remaining plants were like (0. 

 Lamar ckianaXO. biennis Chicago) laeta and velutina. The recipro- 

 cal cross yielded 59 specimens, of which 5 remained in the condition 



" On these questions see Gruppenweise Artbildung, pp. 268-295. 1913. 



