94 THE PHENOMENON OF 



three quinate circles of stamens, which though i prosen- 

 thesis are placed in a ternary relation of alternation, while 

 the abortive circle still belongs to the binary condition of 

 alternation (though i prosenthesis). In the Crassulaceae 

 likewise, the abortive circle may be regarded as an inner 

 corolla, although the arguments for this are less direct. 

 On the contrary, the abortive circle is in some cases 

 decidedly an outer circle of stamens, for instance, among 

 the Monocotyledons, in the Burmanniaceae and Ponte- 

 deraceae, among the Dicotyledons, in many Malvaceae 

 and Tiliaceae, e.g., Helicteres, Hermannia, moreover in 

 the Ampelideae and Rhamneae. Both united, i. e., 

 two abortive circles, a suppressed inner corolla and a 

 suppressed outer circle of stamens, at the same time, 

 occur in the Crassulaceae which have only one circle of 

 stamens properly developed, as for instance Crassula 

 and Tillfsa ; likewise in Erodium, of the family of the 

 Geraniaceae, and the allied Linum ; in Siberia and 

 Azalea, of the heath family.* As we have seen in the 

 conditions of the transition to the flower, examined above, 

 that the whole of the preceding hypsophyllary formation 

 frequently vanishes, so this is repeated here. Not 

 merely particular segments, but even the entire corolla 

 may vanish. We have already seen an approximation to 

 this in the exceptional cases, in which the maximum of 

 the leaf-formation within the flower is not attained in 

 the corolla, but already in the calyx, while the petals 

 appear small and imperfect. t The apetalous flowers 

 produced in this way occur in very many families,! and 

 allow their true nature to be readily decided when we 



* In this and similar cases no circle seems to be absent, since when two 

 succeeding circles are suppressed, the alternation of the properly developed 

 circle is restored. 



f See pages 78, 79. 



\ To prevent misunderstandings, I must remark that the condition of 

 abortion is not the explanation of all apetalous flowers. There are families 

 in which the metamorphosis actually progresses directly from the calyx- 

 formation to the stamen-formation, without any suppressed intermediate 

 formation corresponding to the corolla ; thus, for instance, in the Polygonese 

 and Laurinese. 



