238 FLOiiA INDICA. [NymphceacecE, 



3. The structnre of the rhizome does not deviate more from the Exogenous type 

 than that of many other Dicotyledons. It does not belong to the Endogenous type, 

 and no Monocotyledon is known to have a similar rhizome. 



4. The venation of the leaves is reticulated, and their vernation is involute. 



B. The floral organs are generally arranged upon a quaternary or quinary plan. 



6. TSymplueaceie. present many direct affinities vpith both apocarpous and syncar- 

 pous Thalamiflora, as Uanunculacees, Berheridece, MagnoliacecE, and FapaveracecSf 

 and they present no affinity whatever with any Monocotyledonous Orders. 



7. Systematic botanists are almost unanimously inclined to the above view of their 

 immediate affinities. 



There are very many interesting aud curious points in the structure of Nym- 

 phteacete quite apart from those we have dwelt upon, for which we must again refer 

 to Trecul aud Plauchon, confining our attention to such only as have a systematic 

 value. The floral envelopes usually form an uuiaterrupted spiral from the sepals to 

 the inner stamens, the transition being gradual from one class of organs to the other, 

 as in Magnoliaceai. In Nymphixea the prevalent numbers are four sepals, succeeded 

 by several whorls of eight petals, four opposite to and four alternate with the sepals, 

 and the stamens are similarly disposed ; but in some American species the eight 

 leaves of each whorl of stamens and petals are all opposite one another : this arrange- 

 ment of parts is eminently characteristic of the allied Orders Menispermacea, Berie- 

 ridecp, SabiaoeiE, and Lardizabalets, 



The disc or torus of Nymphaacea is a most remarkable modification of the bases 

 of the periauthial leaves and apex of the peduncle. "We cannot agree with Trecul iu 

 denying the presence of a disc, though it is difficult to assign its limits and origin. 

 The fact dwelt upon by that author, that iu the earliest state of development of the 

 flower, when the stamens and carpels appear as mere points, there is no space be- 

 tween the latter, appears to us to have no weight in this ease, for the carpels are 

 eongenitally imbedded in it, and it appears adherent to the walls of the ovary as 

 these are developed ; it is not a free organ, like the perigynous ring of Alsinefs, and 

 does not arrive at its full development till the floral organs are fully formed. Its 

 structure was flrst clearly explained to us by Bentham, who has shown us that in 

 the fuUy formed fruit of all Nymphcsts the carpels are imbedded in the disc, which 

 rises in the centre of the compound ovary in the form of a cone or mamilla. The 

 ovaries are hence gynobasie. The stamens are inserted into the disc at the base of 

 the ovaries, or all round the whorl of carpels ; or in Victoria the disc is carried up • 

 above the carpels, forming a ring upon which the stamens and petals are inserted. In 

 Cahombeai there is no disc, the carpels are free, and the stamens hypogynous. In 

 Barclaya the four sepals are inserted at the base of the flower, and the petals and 

 stamens carried up upon the disc, which is adherent with the carpels to their summit, 

 whence the calyx is inferior and the corolla superior, as in some species of the curious 

 Himalayan geuus Codonopsls of Cayn^anulaceiV . Lastly, in Earyale and Victoria 

 the whole perianth is superior, which may perhaps be explained by supposing the 

 flower to be sunk in the expanded apex of the peduncle, as in Tiom and perhaps the 

 Pomacece, and to which there is a tendency in Eschscholtzia amongst Papaveracea. 



Between the stamens and carpels there are iu Nymphtsa organs that have been re- 

 garded as incomplete stamina, as appendices to the stigmata, or as prolongations of 

 the stigmata themselves. These are always opposite to the stigmatic lines, and are 

 continuous with the disc below, so that their real nature is not apparent in some 

 eases ; they appear in some to be rudimentary stamina, as in the American Nymphcsce 

 of the hlanda group, though the stigmatic surface is prolonged on to their bases. In 

 N. Lotus they are very large, and are generally regarded as stigmatic appendages ; 

 in N. carulea and its allies they form short horns to the stigmatic rays, and can 

 only be theoretically, if at all, attributed to the presence of rudimentary stamina ; 

 they may be analogous to the stigmatic appendages of EschschoUxia and Fuinariaceie, 

 or to the appendages to the carpels of some other Papaneracete. This point wants 

 a systematic study. 



The fact of the placenta being spread over the whole surface of tlic caviiy of the 



