358 THE JOURNAL OF BOTANY 



numerous bud-bearing internodes above the open flowers and 

 large bracts, and its dark green, shiny, moderately broad leaves, 

 thus answering to the definition of De Vries's CE. scintillans. As 

 the stigma is sometimes well above the anthers, I cannot see how 

 CE. scintillans is to be distinguished from such examples of 

 CE. biennis. The "grossere Gestalt" of the flowers {Mutations- 

 theorie, ii. p. 477) is no criterion, since in a flower of CE. scintillans 

 sent to me by Prof. De Vries the petals measure only 22 millim., 

 whilst they measure 27 millim. in specimens received from him 

 as CE. biennis. How is a systematist to carry out his work in 

 presence of such contradictions ? 



The study of the above examples entirely confirms the opinion 

 expressed by Mgr. L6veill6 (Mon. Onothera, i. pp. 19 and 106) : 



" Des a present nous ferons une remarque fort importante sur 

 laquelle nous appelons toute l'attention des botanistes, c'est la 

 variability de la grandeur de la fleur chez les Onotheres. Ce 

 caract^re se rencontre a peu pres chez toutes les especes qui 

 presentent une forme a grande fleur et une forme a petites fleurs. 

 On ne saurait considerer la grandeur ou la petitesse des fleurs 

 comme des caracteres susceptibles d'etre employes dans la classi- 

 fication, de telle sorte que Ton ne saurait y voir qu'un caract^re 

 accidental r6pondant tout au plus a l'idee de variation. 



11 Chez les Onothera la grandeur des fleurs et la presence de 

 poils ne peuvent fournir un caract6re s6rieux pour 6tablir une 

 esp6ce, pas m6me pour une variete stable. II est impossible de 

 savoir ou finit YO. fruticosa et ou commence la pumila. On trouve 

 chez la pumila des styles 6galant ou meme depassant les 6tamines 

 tout aussi bien que des styles plus courts que celles-ci. . . . 

 D'ailleurs, ainsi que nous l'avons dej& dit, chez toutes les espfeces 

 d' Onothera, on trouve la forme a grandes fleurs et la forme a 

 petites fleurs et sou vent des intermediates entre les deux. Ce sont 

 Ik des variations sans importance que nous nous refusons a 

 enregistrer a titre de variety." 



From what I have seen in the CE. biennis growing in South 

 Kensington I can fully endorse what has been said by the distin- 

 guished French botanist, and I am further strengthened in this 

 opinion by the observations I have made in Brittany. 



II. — Observations in Brittany. 



In 1899 I was spending my summer holiday at a small seaside 

 place on the north coast of Brittany, La Garde St. Cast (C6tes-du- 

 Nord). Close to the hotel there were a few plants of a large- 

 flowered CEnothera. answering to the form often cultivated in 



„ is CE. Lamarckiana or CE. biennis 



grandiflora. Their origin is unknown to the proprietor of the 

 hotel ; they are not escaped from his garden. I have been to that 

 place several times since, and found the plant gradually spreading 

 over the sandhills in the neighbourhood of the hotel. 



In 1904 I had with me De Vries's famous book Mutations- 

 theorie, and I endeavoured to determine the varieties of this 



