CBUGIFERJE. 213 



as in Sisymbrium, Nasturtium, Lnarthrdcarpus, Streptant/tus, Sgrenia, 

 Vella, Boleum, Nothothlaspi, &c. When it is proved that in this order 

 the secondary axes of the inflorescence are often carried up to a 

 variable height on the chief axis, it will easily be seen that they 

 really spring from the axil of a leaf or bract, but that they separate 

 from the main axis so far above as to obscure their true relations 

 when adult. 



8. The floral receptacle is nearly always convex, usually forming 

 a depressed cone ; this involves a hypogynous perianth and gyna> 

 ceum. But in one series (the Subulariece), as we have elsewhere 

 noticed particularly, 1 the receptacle forms a pretty deep cup, on the 

 edges of which the stamens and perianth are perigynous. 



9. The usually hypogynous glands 2 internal to the corolla, which 

 have been considered by many authors the representatives of more 

 or less abortive floral appendages, especially stamens when they are 

 long and narrow, 3 appear to us to be, as Adanson conjectured, 4 depen- 

 dencies of the receptacle, regions thereof hypertrophied late, as is the 

 case with disks generally. The peculiar forms of these bodies, and the 

 limited spots of the receptacle on which they form, would seem to 

 depend on the free spaces left for their growth by the true floral ap- 

 pendages between which they occur. This may be easily seen in certain 

 genera wherein the receptacular surface is completely lined by a (yellow- 

 ish or greenish) glandular layer, except at the depressions for insertion 

 of pistil and stamens. In others the glandular surface is divided into 

 two islets on opposite sides of the flower, either right or left, or, 

 more rarely, in front and behind. Yet more frequently are there 

 four : either two anterior and two posterior, one in front of each 

 petal ; or oftener, two lateral, and two antero-posterior. In the 

 latter case they embrace completely or partially a small stamen, 5 or 

 two large ones, respectively. 6 Very often, too, the hypertrophy is 

 localized inside the two small stamens, and outside the large ones. 

 Jn such cases especially are the glands developed into crescents, 

 scales, horns, or spurs ; and their form is so variable as to induce 



1 See Adansonia, x. 45, t. 6. 4 See Fam. des PL, ii. 412. 



2 See Pater, Organog., 213. 5 These glands are termed carpellary, like the 



3 In this case the gland outside the pair of two lateral sepals, because of their position, 

 large stamens has been held a staininode belong- 6 Similarly, these are termed placentary. 

 ing to the same verticil as the two Literal small 



stamens. 



