AD FLOEAM INDICAM — 0AMPANULACEJ3. q 



stem and in the axils of the leaves, than are ever found in the 

 Indian plant. 



Under Oampanulacece we have included the Loleliacece as a tribe, 

 considering that they have no sufficient claims to rank as a separate 

 order. It is true that the limits between these tribes are seldom 

 disputed ; but they are of very little value, and are founded on 

 characters common to many species of both. The character most 

 relied on is the connate anthers of Loheliacece, but these are found 

 in Symphyandra of Campanulacece, whilst various LobeliacecB have 

 free anthers. Even the irregular corolla affords no good mark, 

 for some states of the Australian Wahlenhergia saxicola have an 

 oblique corolla, and unequal inclined anthers, of which two have 

 the connective produced into an appendix, thus imitating a pre- 

 valent feature of LobeliacecB. In both tribes the fruit is either 

 baccate or capsular, and in each the dehiscence takes place some- 

 times above and sometimes below the limb of the calyx. 



The Indian Loheliacece offer few structural peculiarities. "We 

 have, however, been obliged to found a new genus upon a remark- 

 able and handsome species from the Sikkim Himalaya, identical 

 with a mountain plant of Java, which we suppose to be the Lobelia 

 montana of Blume. 



The Indian Campanulece are a more extensive and very in- 

 structive tribe, and we have several remarkable forms to add to 

 those which have been so well illustrated by Alphonse DeCandoUe 

 in his elaborate and able monograph. Of these novelties Codo- 

 nopsis and its allies are the most anomalous, including as they do 

 the beautiful genus CyanantJius, which, following Bentham, we 

 unhesitatingly refer to this group, and which is indeed scarcely 

 separable by technical characters from Wahlenhergia itself. In 

 the ' Illustrations of Sikkim Himalayan Plants,' we figured the 

 most remarkable Indian forms of Codonopsis, regarding several of 

 them as subgenera, though founded upon extreme deviations from 

 the prevalent arrangement of the floral whorls in Campanulacece. 

 Our friend M. DeCandoUe has since communicated to us by letter 

 some valuable criticisms on the course we adopted ; pointing out 

 that some of the characters which we held to be of only subgeneric 

 value, are of even ordinal value in other families of plants ; and 

 further, that if these subgenera are to be permanently considered 

 as such, almost all the genera of Campanulacece may be merged 

 into one. Of this we were perfectly aware, and indeed of more 

 than this ; namely, that the whole question of what should or 

 should not constitute a genus, is involved in the consideration of 



