362 THE JOURNAL OF BOTANY 



leaves (as in CE. muricata L. and CE. elliptica De Vries), and these 

 are most conspicuous for that character, even at a distance (Nos. 26 

 and 32), but then, if we abide by the pistil standard of De Vries, 

 some would be referred to the CE. biennis group, others to the 

 (E. Lamarchiana group. No. 2, as regards the leaves, connects 

 this variety with CE. Lamarchiana. 



I must point out that the plants which bore flowers (Nos. 25 

 and 27) so different in the length of the pistil, did not differ 



appreciably from each other. 



In short, I have tried hard to refer all these variations to the 

 14 elementary species" of De Vries, but without success. Acting 

 on the principles advocated by De Vries, it would be possible to 

 describe a further number of new species, in addition to recognizing 

 (E. biennis and CE. Lamarchiana in the restricted sense, but their 

 definition would be very shaky. 



As I feel tolerably certain that no Oenotheras have been intro- 

 duced in the locality since the few large-flowered plants were seen 

 by me in 1899, I have come to the conclusion that, at St. Cast, 

 the curious forms which have appeared, as the plant has spread, 

 are all descendants of CE. Lamarchiana, which, when left to itself, 

 may in certain localities, besides producing other variations, to 

 be further multiplied by crossing, revert to a form so closely 

 similar to the true (E. biennis, from which it is probably derived, 

 as not to be distinguishable from it. 



If some of the mutants can be shown to be merely reversions 

 to an ancestral type, how can the others be looked upon as 

 species ? They also may be reversions to some type out of which 

 a form of obscure origin like CE. Lamarchiana was first evolved, 

 possibly by repeated crosses between varieties of the polymorphic 

 CE. biennis, which would be quite in accordance with the results 

 of Louis de Vilmorin and Naudin. 



Since my return from Brittany I have read Mr. Charles Bailey's 

 paper on " The Evening Primrose on the Sandhills of St. Anne's- 

 on-the-Sea, North Lancashire" (Manchester, 1907), from which I 

 see he also has experienced difficulties in reconciling his plants 

 with the definitions given by De Vries. He found variations in 

 the petals and pistil, as well as in the leaves, some plants having 

 the midribs of a pinkish red or light crimson colour, but this 

 character of (E. rubrinervis was not associated with the brittleness 

 issigned to that species, or with red streaks on the capsule. He 

 also mentions plants, by no means rare, with lighter yellow flowers 

 than Lamarchiana, and only half their size. In one of these forms 

 with small flowers the stigma was longer than usual, and its four 

 divisions lay at the base of the corolla, the style being nearly 

 suppressed, the anthers showing fully half an inch above the pistil ; 

 this form Mr. Bailey takes to represent De Vries's brevistylis. In 

 the other small-flowered form the stigma and anthers were of 

 equal height. This latter form, I should say, represents the 

 typical (E. biennis, to which CE. Lamarchiana may revert. It 

 seems, however, to judge from Mr. Bailey's remarks, that the 

 Evening Primrose does not exhibit on the sandhills of St. Anne's 



