96 THE PHENOMENON OF 



inner stamens are longer than the three outer, also in 

 the Cruciferse, in the well-known tetradynamous con- 

 dition ; in Asarum, with six outer shorter, and six inner 

 longer stamens ; in Bheum, with six outer shorter, and 

 three inner longer ; in Oxalis and Limnanthes, with five 

 outer shorter, and five inner longer ; in Monsonia, with 

 ten outer shorter, and five inner longer; lastly, most 

 beautifully in Hibiscus and other Malvaceae, in which 

 numerous quinate circles of stamens are piled up into a 

 more or less abundantly clothed column. Most of the 

 Ranunculacese, in particular Anemone decidedly, exhibit 

 a gradual increase of length of the stamens, together 

 with an acyclical, spiral arrangement of them ; and many 

 of them a decrease again at the close of the formation, 

 as in Pceonia Montan, the innermost, shortened, and 

 abortive staminal leaves of which, form by their con- 

 fluence the well-known crimson crown round the germen, 

 which DecandoUe considered as one of the principal 

 supports of his theory of the torus.* 



The structure of the Chinese Peony just noticed, leads 

 us to the examination of the transition from the stamen- 

 formation to the fruit, the last chasm which the plant 

 has to pass over in its path to the goal. In this transition 

 to the last stage of the metamorphosis, a complete sup- 

 pression of the leaf-formation occurs far more frequently 

 than at the transition from the corolla to the stamen- 

 formation, a circumstance of especial importance for a 

 correct insight into the structure of most flowers. In the 

 Monocotyledons only, in which in general the formations 

 are less sharply separated, a direct transition, uninter- 

 rupted by intermediate abortive circles, is the more 

 frequent condition, while in the Dicotyledons this occurs 

 as the rule only in few famihes, particularly such as 

 exhibit aflinity to the Monocotyledons in other respects ; 

 in many others not at all, or only in isolated genera. 

 This also furnishes the explanation of the rarity of 



* Decandollc, ' Organographie,' i, 181. 



