18 D. Prain — Some additional Fnmsirisicess. [No.], 



Since 1872, when the account of this genns in the Flora of British India was 

 published, the number of species reported from the Indian area has been doubled. 

 For our acquaintance with eight of the newly reported species from the north- 

 west Himalayan region, we are indebted to the exertions of Mr. Duthie of Saharan- 

 pur; ten more are due to the extensive exploration of the provinces of Sikkim and 

 Chumbi conducted by Dr. King daring the past twenty years. Four others from 

 the north-west and north-east frontiers have been obtained by Dr. Aitchison, Mr. 

 Ellis, Mr. Lace and Dr. Watt. The remaining species, recognised in this paper as 

 separate, are plants known at the time of publication of the Flora of British India, 

 but in that work referred to other species. In nearly every case they had already 

 received the rank of varieties ; in every instance not only their existence but the fact 

 that they exliibit characters deviating from those of the species to which they have 

 been tentatively referred, has been pointed out by Sir Joseph Hooker and Dr. 

 Thomson ; their present recognition as species apart is due to the communication 

 since 1872 of more extensive material for study. 



In the foregoing Key, which has been prepared principally with a view to the 

 assistance of field botanists, pains has been taken to avoid as far as possible the use 

 of floral characters. The characters derived from the flowers do not in the writers' 

 experience assist one greatly in classification. The relative length of spur and 

 lamina of the larger outer petal is not quite a reliable character; though in the 

 majority of cases this i-elationship remains fairly uniform, there are some in which 

 it does not, there being considerable variability in the absolute length of spur 

 within the limits of some at least of the species here recognised, without any cor- 

 responding altei-atiou of dimensions on the part of the lamina. The presence or 

 absence of wings to the petals is another character that, taken by ilself, appears to 

 fail; at all events among Indian species it has been found necessary to include in at 

 least two, G. cachemiriana and G tibetica. that are widely divergent, forms which 

 save for the complete absence of wings to the petals cannot be distinguished 

 from their respective types. Nor is colour of material assistance. In the large 

 majority of Indian species the flowers are some shade of yellow, in one instance 

 (C. ophiocarpa) so faint that the flowers are almost white; in the remaining 

 species the flowers arc mauve or purple. But one species with usually purple 

 flowers (C fiaccida) sometimes has yellow petals, and two species usually with yel- 

 low flowers (C. meifolia and G. crispa) sometimes have them mauve. 



That good characters for purposes of classification are likely to be obtained 

 from the fruit and seed is very probable. But in a considerable number of cases ripe 

 fruits and seeds are still unknown ; it is very diSicult to obtain the fruits of autumn 

 flowering species owing to the necessity that collectors are under of hurrying away, 

 before their fruits are fully ripe, from the inclement altitudes that many of the 

 species affect. In the meantime, therefore, it has been deemed advisable to use 

 for purposes of arrangement and, as far as possible of specific diagnosis also, the 

 more general characters derived from habit and foliage. 



It may be mentioned that it has been found impossible in drawing up the 

 Key to retain unbroken the section of species with fibrous roots and 1-seriate 

 seeds proposed in the Flora of British India. One of the three species included in 

 that section proves to be possessed of a fusiform rootstock and to be more nearly 

 allied to C. jiahellata and G. adiantifolia, two species with also 1-seriate seeds though 

 ah-eady placed in the other section. On the other hand G. Laelia, a new species from 

 Sikkira, has fascicled fibrous roots and oval capsules with 2-seriate seeds. The sec- 



