264 REMARKS ON DOORNIA AND RYKIA, 



m 



conclude, that the hitherto known Java forms should be separated from 

 the genus Pandanus. This is obvious from what Kuntb has advanced 

 in his ' Enumeratio ' (iii. 1841, p, 94), concerning the plants which he 

 attributes to the genus Pandanus, He takes up thirty sorts, of which 

 scarcely one, viz. P. odoratmimus (and this not in all respects) can be 

 received as known. It is rather to be called a list, or catalogue, which 

 in many points has no more value than a gardener's list, so short are 

 the descriptions of the species represented. The investigations of Bory 

 de St. Yincent, and Aubert du Petit-Thouars, have indicated many sorts 

 as growing on the Isle of Trance, Bourbon, and Madagascar; but from 

 which enumeration it appears that we cannot sufficiently recommend re- 

 searches in these islands, for of these plants we do not as yet know any- 

 thing really. We are equally ignorant of the Australian species. The 

 two sorts which Brown has mentioned, are only briefly described. 



If we are really to advance in our knowledge of this family of plants, 

 we must have coloured drawings of the plants in their natural state, with 

 full descriptions ; the flowers and fruit preserved in spirits ; collect the 

 dried leaves, describe them on the spot, or, if possible, send them to 

 Europe. ■ That this is particularly necessary with regard to the female 

 plant and the ovule, is obvious, from the beautiful work of Gaudichaud, 

 who, in the botanical part of his Yoyage round the World in the corvette 

 La Bonite, performed in 1835-7, shows the very plan to be adopted by 

 those who would investigate the Pandani, Though many calm inves- 

 tigators may disapprove the far-stretched hypotheses of this botanist on 

 the growth of the monocotyledonous stem, it is however admitted, by 

 all who can value his labours, that Gaudichaud is one of the most emi- 

 nent analytics in modern science. Also, in respect oi Pandanece^ the re- 

 searches of Gaudichaud are excellent, although only known by drawin 

 to which no text is annexed. In his excellent analyses accompanying the 

 plates in his admirable atlas of the above-mentioned work,M. Gaudichaud 

 has enumerated the following new genera: — Barrotia (pi. 13), Bryan- 

 tia (pi. 20), Dorydlgma (pi. 13, 31), Fiscpietia (pi. 4), Vimonia (pi. 17, 

 23, 31), Roiminia (pi. 21), Homhronia (pi. 20), Sussea (pi. 24, 25, 38), 

 Jeannerettia (pi. 25), Heterostigma (pi. 25), Toullioya (pi. 26), Tucleya 

 (pi. 26), Eydouxia (pi. 18), Souleyetla (pi. 19). Ten species of Pan- 

 dani are proposed analytically. Among them, as among the new 

 genera, arc included, naturally, more of those of the old sorts : — P. titilis 

 and sylveslris (a rinsonid), P. edulis (Homhronia), P.conoidem {Sassea)^ 



S^ 



