Structure of Flowers of Adoxa moschatellina. 261 



in the filaments to combine in pairs. Following up the hint which 

 was thus afforded, I found that the anthers were composed of a single 

 cell ; and that it was in those cases only where two contiguous fila- 

 ments had become completely united, that we ever have a perfect sta- 

 men crowned by a two-lobed anther. This at once solved the difficulty, 

 and reduced the structure of the flower within the usual conditions. Dr 

 Hooker has not decided which structure should be adopted, and has 

 followed the arrangement generally adopted of classing this plant 

 under Octandria. The fact of the frequent coherence of the filaments 

 in the contiguous stamens, combined with the regularity thus introdu- 

 ced into the arrangement of the several parts of the contiguous whorls, 

 singularly strengthens the conjecture he had hazarded from the con- 

 sideration of the anthers being one-celled, and may indeed be con- 

 sidered as affording as direct a proof of the normal condition of this 

 flower as the subject will admit. It may therefore be stated as con- 

 taining 4 whorls of 5 parts each, the parts alternating in the suc- 

 cessive whorls, — a structure eminently characteristic of a dicotyle- 

 donous plant, and probably exhibited in greater perfection in the 

 genus Crassula than in any other. In the present case, the devia- 

 tion from the normal character in the lateral flowers consists in the 

 ordinary suppression of 2 or 3 parts in the calyx, sometimes of 1 

 only ; and in the terminal flower in the suppression of one part in 

 all the whorls as they are usually exhibited in the lateral flowers. 



Among the numerous specimens which I 

 examined, some had the stamens partially con- 

 verted to leaves ; two small ones being placed 

 facing each other and occupying their usual 

 position on the corolla. (Fig. 4.) In some 

 cases a supernumerary petal of small dimen- 

 sions was accompanied by a thick filament 

 bearing a three-lobed anther, a monstrosity 

 which apparently originat- Fig. 5. Fig. 6. 



ed in a complete and super- 

 numerary stamen having 

 been also developed, and s 

 united with the half sta- 

 men to which it was con- 

 tiguous. In other cases, 

 the contiguous filaments 

 were united, and one part foliaceous, the other 

 antheriferous, recalling the structure of the sta- 

 mens in a Canna. 



1 



Fig. 7. 



