ON THE VARIATIONS OF THE EVENING PRIMROSE 359 



CEnothera with the aid of the descriptions and figures it contains. 

 But I did not succeed, and felt highly perplexed at the great 

 amount of variation I found among these plants, some of which 

 agreed in the large size of the flowers and the very long pistil with 

 the general definition of (E. Lamarckiana, whilst others came very 

 close to (E. biennis. But so numerous and bewildering were the 

 intermediate forms that I gave up the task in despair. Yet I 

 felt absolutely convinced that all these aberrant specimens were 

 descended from the few plants which grew there in 1899. I came 

 to the conclusion that, having run wild, these Oenotheras were 



(E 



other forms similar to 



Mutationstheo 



However, I found it difficult to marshal them under the defini- 

 tions given by De Vries, and I felt serious doubt as to the value of 

 the characters adduced by him in support of his " elementary 

 species," seeing that the flowers at least occasionally varied 

 on the same plant beyond the limits assigned by him to his 



" species " 



wn 



Last August I went back to La Garde St. Cast, and found the 

 Oenotheras had further spread. Although the specimens were not 

 so numerous as I should have wished, owing to people having 

 picked many, probably the finest of the inflorescences, I examined 

 a large number of them, and was again struck by the enor- 

 mous variation they exhibited in size and in the shape of the 

 various organs. I dried a selected series of specimens, which I 

 have deposited in the Natural History Museum. I have examined 

 them again, and feel more convinced than ever of the futility of 

 attempting to define species among them. It would be also quite 

 impossible to divide them into two groups — (E. Lamarckiana and 

 (E. biennis — according to the standard set forth by De Vries, viz. 

 the size of the flowers and length of the pistil compared to that 

 of the stamen. 



The first thing that strikes one is the variation in the size 

 of the petals and in the length of the pistil. This is sh< 

 by the following table, in which I have arranged the various 

 specimens according to the size of the flowers, irrespective of 

 other characters to which I shall allude further on, the arrange- 

 ment being the same as in dealing with the South Kensington 

 flowers. All the flowers here recorded are from different plants, 

 with the exception of three (marked *) which are from the same 

 inflorescence, their measurements being given to show that con- 

 siderable variation in the length of the pistil may exist in the 

 same plant : — 



A. B. C D. 



1. 48 26 22 38 



2. 46 33 20 40 



3. 45 37 18 35 



4. 43 32 19 34 



5. 43 30 18 40 

 • 6. 42 32 - 20 42 



7. 42 25 17 40 



8. 41 30 18 42 



