26o BOTANICAL GAZETTE [October 



The remaining 4-7 and 3 per cent were mostly of the Lamarck- 

 iana type, with some mutants belonging to alhida, oblonga, and 

 nanella. I had saved the fruits and their seeds in 7 or 8 groups, 

 beginning at the base of the spike, and sown the seeds separately, 

 But, just as in the previous case, there were no appreciable differ- 

 ences in the percentage figures between the higher and the lower 

 groups. 



The main result is that the percentage of specimens of the 

 cana type, which runs 15-60 per cent on annual individuals, 

 may increase to 93-97 per cent on very vigorous biennial plants. 

 It is thus clearly seen to be dependent upon the method of culti- 

 vation. Obviously this rule may be appKed to the percentages of 

 O. scintillans, as previously discussed, and to those of 0. pallescens 

 and the other new dimoiphic mutants to be described in this article. 



Oenothera Lamarckiana mut. pallescens (fig. 3). — ^Among all 

 the mutants which arose in my garden from O. Lamarckiana, this 

 form most closely resembles the parent tj^e. In early stages the 

 rosettes are the same, and in springtime, when still in the boxes, 

 I have not as yet succeeded in distinguishing them. It is not 

 until about 6 weeks after planting out on the beds that the differ- 

 entiating marks begin to show (fig. 4). In the middle of June the 

 leaves are clearly shorter, and the blade is set off from the narrowly- 

 winged petiole by a sharp indentation. This character causes the 

 rosettes to be more open because the petioles hardly touch one 

 another. 



This spatulate form of the leaves remains, for a long time, the 

 best mark of the race; but when the stem grows up, the whole 

 plant is much more slender than the parent form (fig. 3). The 

 stem is thin and low; in July, when the first flowers open, it often 

 reaches only 75 cm., when the corresponding specimens of Lamarck- 

 iana are already i m. and more in height. After a time, however, 

 this difference disappears, since the spike is more elongated. It 

 is less dense than in Lamarckiana; the bracts are much shorter and 

 strikingly broader; the flower buds are large and conical, the 

 flowers somewhat smaller, although still larger than those of 

 0. biennis; the pollen is abundant and the fruits are short and thick, 

 containing a good supply of seed. The foliage is of the same green 



