POSITIVE DifiDOUBLEMENT. 79 



together represented a single organ, and hence the 

 flower possessed but three alternating whorls alto- 

 gether. But here the usual mistake was made of 

 relying on the misleading data of ontogeny. Eichler 

 sets forth the correct view which, as always, is gained 

 by the comparative study of forms. In other members 

 of the Primulacese, e.q. the brook-weed (Samuhis) and 

 the yellow loose-strife {Lysimachia vulgaris), there 

 occur, alternating with the stamens and petals, small 

 scales, glands, or teeth which Braun regarded as a 

 second corolla, but which are usually considered as 

 pertaining to the androecium, and hence called stami- 

 nodes. In the allied orders Myrsinaceee and Sapotacese 



■/Dei- 



/^eZ: 



Fig. 92. — Dianthus coronarius. Serial dedoublement of stamen-rudiments' 

 to form a "double-flower." pet, normal petals. (After Goebel.) 



these structures are often better developed, and bear 

 either vestiges of anthers or are complete fertile 

 stamens (as in Isonandra). Moreover, in a flower of 

 the primrose (P. vulgaris) there was an outer whorl 

 of stamens alternating with the petals and normal 

 stamens. Hence we must p,ssume that in the normal 

 flower of Primulacese and Plumbagineae the outer whorl 

 of stamens has become completely suppressed and the 

 inner whorl has congenitally fused with the corolla. 

 In P. Auricula Marchand observed a third whorl within 

 the normal whorl. 



In the Iridaceae it is the stamens of the inner whorl 

 which have disappeared, so that the stamens which are 

 present come to be opposite the carpels : an otherwise 



