i9i6] devries—dimorphic mutants 267 



type remained small, their leaves reaching only about half the 

 length and half the breadth of those of their sisters (7X3 cm. 

 against 14X5 cm.)- Their blades were sharply set off from the 

 winged petioles (fig. 5, C), and thereby they much resembled those 

 of O. pallescens (fig. 4). This resemblance continued during the 

 development of the stem and the flower spikes; but even as the 

 rosettes were smaller and more densely leaved, the stems were 

 lower and weaker and less branched. The leaves were narrower 

 and folded along the middle vein, instead of being broad and flat- 

 tened, as in 0. pallescens. The flower buds were as thick and as 

 large as those of this species, and the flowers also reached the same 

 size, the petals having a length of 4 cm. During the flowering 

 period the differences from 0. pallescens grew gradually less, and 

 at the end, in September, the new type seemed to be only a weak 

 form of this latter, reaching a less height and being almost un- 

 branched. Artificial self-fertilization has been difficult, since in 

 many flowers the pollen was in an imperfect condition. Four 

 specimens yielded a sufficient harvest (0.5-icc.). The sister 

 plants of the Lamarckiana type showed an abundance of seed,, 

 exactly as the Lamarckiana of pure origin does. 



The next year (1915) I sowed the seeds of one specimen of the 

 parental type and of two of the type of Lamarckiana. The first 

 gave only 44 seedlings, of which 4 were Lactuca, one nanel.a, and 

 the others Lamarckiana. The two other sowings gave 248 and 283 

 offspring of their own type, without any Lactuca specimens, but 

 with some dwarfs. The number of them was 1 1, or 4 per cent in the 

 first group, but only one in the latter group. Combining these 

 results we get the following pedigree : 

 1913 Mutant Mutant 



igi4 First generation 43 per cent Lactuca Lamarckiana 



1915 Second generation 9 per cent Lactuca Lamarckiana ■ Lamarckiana 



Although only two generations from the seeds of the original 

 mutant have been cultivated, it is evident that this new form 

 behaves exactly like the inconstant races of 0. scintillans, 0. cana, 

 and 0. pallescens. Under favorable conditions it splits into about 



