Corydalis.'] floua indica. 239 



described, in which the majority of the species are upon the whole remarkably well 

 marked and distinct from one another ; amongst the Himalayan ones, at any rate, 

 there is none of that interlacement of forms that has rendered the disentanglement of 

 the species of Banunculacere and Berberidem so laborious and unsatisfactory. 



We have not adopted the sectional groups proposed by De CandoUe, as they do not 

 seem to be altogether natural, and some of the best characters by which they are 

 limited (those of the root, for instance) are practically unavailable. Many of the 

 species have tuberous roots, but in a considerable number these are so deeply buried 

 in the earth or lodged in crevices of rocks, that it is impossible to prove their 

 existence in the living plant. A knowledge of the roots of the species is a great 

 desideratum, which we often in vain attempted to supply, and the more to be re- 

 gretted because the characters they afford are eminently natural. With regard 

 to the character taken from the length of the spur of the posticous bundle of 

 filaments, that seems to depend mainly upon the length of the spur of the pos- 

 ticous petal itself ; and where it does not, a strict adhesion to its proportional length 

 would sunder very closely allied species. The persistence of the style is a very in- 

 constant character, and that drawn from the lobing of the stigmata is not available 

 in dried specimens, and of doubtful value. The arillus varies extremely in form and 

 relative size during different stages of the growth of the seed, and is not quite con- 

 stant in each species. A much more important character is drawn from the develop- 

 ment of the young plant ; the seed in the section Bulbocapnos being described as ger- 

 minating by a single cotyledon, whose radicle forms a perennial tuber, which sends up 

 a primordial leaf in the following year, and a flowering stem in subsequent ones : the 

 other sections, again, have opposite cotyledonary leaves. It is evident, however, that it 

 must be many years before observations on this point can be verified on even a few 

 species of the genus, and until done for the majority, the value of the characters they 

 afford must be quite problematical. Lastly, the sections Capnoides and Ca/pnites are 

 hardly distinguishable by any character, and we find species placed in each that should 

 certainly stand veiy close together. Under these circumstances we have not hesitated 

 to take definite characters drawn from the pod for the primary divisions, and others 

 from the perianth, etc., for those of secondary value. These, however, are in a great 

 measure arbitrary, and are proposed as provisional only. 



The maximum of the genus Corydalis is certainly to be sought in the Himalaya, 

 where the species of the western mountains differ so much from those of the eastern, 

 that there are no doubt others to be discovered, especially in Bhotan, Abor, and 

 Mishmi. In the mountains of western China also they probably abound, and there 

 are a considerable number of known but nndescribed species even in the eastern and 

 drier parts of that empire. With the exception of one species, and that a common 

 Himalayan and Siberian one, found in the Khasia, the genus finds its southern limit 

 in the Himalaya. 



Of the 24 species we have described, 9 are new, a much larger proportion than in 

 any other genus hitherto described in this work. In this respect Corydalis is rivalled by 

 very few, except Rhododendron, Impatiens, and Asirayalus. We have also added 2 

 Siberian and 1 European species not hitherto supposed to be Himalayan. Of the 

 Himalayan species 12, or one-half, are found to the eastward of the valley of Nipal, 

 and 7, of which 6 are new, are confined to the eastern Himalaya. On the other hand, 

 16 are found to the westward of the valley of Nipal, of which 10 are confined to 

 the western ranges, and only 3 are new. If, however, we exclude the more strictly 

 Tibetan species of the western regions, some or most of which probably occur iu 

 eastern Tibet also, we have 10 western forms, of which only 4 are not found east of 

 the valley of Nipal. Hence we may infer that the damp regions of the eastern Hi- 

 malaya are the most favourable to the development of species of this beautiful genus. 



Sect. 1. — &7<5'((fl longe lineari-elongata. Sevnina 1-seriata. — Herbse 

 elata ramosa foliosa, mdice Jiii-osa. 

 1. C. ophiocarpa (H.f. et T.); gracilis, ramosa, foliis bipinnati- 



