AD TLOEAM INDICAM — CAMPANULACE^. 7 



' Leptocodon is a third genus, containing one species, entirely 

 agreeing with Codonopsis and Campanumoea in its habit, ramifi- 

 cation, and foetid juice, and with the former genus in the structure 

 and adhesions of its calyx, corolla, &c., but having five staminodia 

 alternating with its stamens, and few ovules fixed to placentaa 

 which are not in the axis of the cells of the ovary, but on the walls 

 of the dissepiments. 



Turning to the fruit of these genera, we usually find in Codo- 

 nopsis a membranous or baccate fruit, bursting by three horny 

 valves within the base of the corolla; but in C. inflata these^ 

 valves appear to be obliterated, and the fleshy berry scatters its 

 seeds by decay of its walls, like that of Campanumoea and Cyclo^ 

 codon. 



In dividing these difficult CampanulaeecB into genera, M. l)eCan- 

 dolle has laid much stress on the relation of the cells of the 

 ovary to the calyx-lobes ; but these, being rarely equal in number, 

 are seldom available for technical characters. In Campanumoea, 

 the sepals alternate with the ovarian cells ; in Cyclocodon, they are 

 opposite ; in Codonopsis rotundifolia (when 5-celled), they are nor- 

 mally opposite, I believe, though I find some observations re- 

 cording the contrary amongst my notes on the live plant. 



Thus in keeping up the four genera^ Campanumoea, Codonopsis, 

 Leptocodon, and Cyclocodon, we find ourselves obliged to go beyond 

 M. DeCandolle, who, in his last revision of the CampanulaeecB 

 (Prodr. vol. vii.), places all the species of Codonopsis, as restricted 

 by us, in Wahlenhergia, and retains the name Codonopsis for 

 "Wallich's C. truncata and C. parviflora, also keeping up Blume's 

 genus Campanumoea. But as one of his species of Campanumoea 

 is identical with Codonopsis parviflora, Wall., and the typical 

 species of Codonopsis differ from Wahlenhergia, it is necessary to 

 remodel all these genera. In the mean time, considering the 

 present state of the genera of Campanulacece, we have thought it 

 better to adopt M. DeCandolle' s opinion, and raise the subgenera 

 of Codonopsis to the rank of genera. 



The only other important observation to which we would call " 

 attention, is the presence in the same individual of some spe- 

 cies of Campanula of two .forms of flowers, differing remarkably, 

 not only in size and appearance, but in structure ; of these one 

 kind is normal, the other very minute, generally on very long 

 and more slender pedicels, with a more globose ovary, differ- 

 ently formed (often more foliaceous) sepals, and no corolla or 

 stamens. These dimorphous flowers sometimes occur on the same 



