266 



much more important consideration than either of these. In true Scitaminea?, 

 as Mr. Brown has observed, (Prodr. 305.) the stamen is always placed oppo- 

 site the labellum or anterior division of the inner series of the corolla, and pro- 

 ceeds from the base of the posterior outer division ; while the sterile stamens, 

 when they exist, are stationed right and left of the labellum. But in Maran- 

 tacese the fertile stamen is on one side of the labellum, occupying the place of 

 one of the lateral sterile stamens of Scitamineee. This peculiarity of arrange- 

 ment indicates a higher degree of irregularity in Marantaceas than in Scitami- 

 neas, which also extends to the other parts of the flower. The suppression of 

 parts takes place in the latter in a symmetrical manner ; the two posterior 

 divisions of the inner series of the perianthium, which are occasionally absent, 

 corresponding with the abortion of the two anterior stamens. In Marantacea?, 

 on the contrary, the suppression of organs takes place with so much irregu- 

 larity, that the relation which the various parts bear to each other is not 

 always apparent : instead of the central stamen being perfect while the two 

 lateral ones are abortive, as in Scitamineee and most Orchidere, or of the cen- 

 tral stamen being abortive and the two lateral ones perfect, as in some Orchi- 

 deae, it is the central and one lateral one that are suppressed in Marantacea?. 

 In the perianthium of Carina only the most external within the calyx can 

 properly be called corolla ; the remainder of the segments being attempts to 

 produce barren petaloid stamens analogous to what is called the inner limb of 

 the corolla in Scitamineee ; and the characters upon which botanists found 

 their specific distinctions depend upon the degree to whch this developement of 

 petaloid abortive stamens extends. When, for instance, they describe some as 

 having an inner limb of 2 or of 3, or of 4 or of 5 segments, they should rather 

 say 2, 3, 4, or 5 stamens are partially developed. For remarks upon the proof 

 thus afforded of the affinity of Scitamineee and Marantaceee to Musaceas, see 

 the former order. 



Perhaps it will be possible to put the relative structure of Scitamineee and 

 Marantaceae in a clearer light by the following diagrams, in which the triangle 

 C, C, C represents the calyx, the angles corresponding with the position of 

 the sepals ; the triangle P, P, P the corolla ; R, r, r an outer series of petaloid 

 stamens, of which r, r are rudimentary only ; and S, s, s the inner series of 

 stamens, of which S is the fertile and fully developed one. 



SCITAMINEEE 



