A NGIOSPERMS. — DICOTYLEDONS. 



457 



so that the sum total of the stamens is no multiple of the typical fundamental number, 

 which in this case is five. Among pentamerous flowers the Lythrarieae, Crassulaceac 

 and Papilionaceae, among tetramerous the G^^nothereae vn.\y be mentioned as 

 exemplifying the interposition of a complete staminal whorl. 



One of the most remarkable deviations from the ordinary conditions is found in 

 several families of the Dicotyledons, in which the single staminal whorl is superposed 



eae. Fir.. 358. Floril dinifrani (,f / VAv (Aiiipcliik-at-). 



on the coroUine whorl, as in Fig. 397, 398, and also in the Rhamneae, Celastrineae, 

 the pentandrous Hypericineae, Tilia and many obdiplostemonous flowers ; Pfeffer ' 

 discovered that the two superposed whorls in the Ampelideae are formed separately 

 and in acropetal order, but that in the Primulaceae they appear as five protuberances, 

 each of which forms a stamen and afterwards produces a petal on its outer side. In 

 these cases there is no sufficient ground for the assumption that an alternating whorl 

 has disappeared between the two superposed whorls ; in the other cases however 

 this assumption is justified or at least very probable. Thus in the scries of Carvo- 



phyilineae families, genera and species occur in which the corolla is wanting, and the 

 stamens are superposed on the sepals ; since in the same alliance there are plants 

 with corollas, we may assume that where these are wanting, they are abortive ^ ; the 

 diagram of these plants is made more complicated by an evident tendency to 

 a doubling of the stamens (Figs. 399, 400) and even of the carpels. 



When there are more stamens in a flower than sepals or petals, this may be 

 the result, as has been already pointed out, of an increase in the number of the 

 staminal whorls, as is shown in Fig. 392, or by interposition of new stamens between 

 those already formed, as in some Rosaceae (Fig. 404), or by an increase in the number 

 of young stamens along with diminution in their size, as in Po/etttilla and Rul>tis 

 (see p. 417), or by chorisis of the stamens, as shown in Fig. 399 ; these cases must 



' Pfeffer, Zur Bliithenentwicklung d. Primulajecn n. Ampelideen in Pringsheim's J.ilirl.. \ III. 

 p. 184. 



2 For another explanation see Kichler, Bluthcidiaijrnnime, II. p. 7S. 



