1896.] D. Prain — Some additional Fumariacere. 19 



tions recognised in the most, recent revision of the natural order* seem already to 

 reqnire reconsideration on account of the enormous accession during recent years of 

 Chinese species; the writer therefore refrains for the present from giving names to, 

 or even from attempting to define, the limits of the more or less natural groups that 

 occur in the genus. It is however only just to those who may consult this Key, to 

 explain that it has been made to adapt itself to as natural a serial arrangement of 

 the Indian species as it has been possible to draw up; in no single instance has a 

 species been intentionally removed from the vicinity of its nearest allies to suit the 

 exigencies or to facilitate the construction of an artificial Key. 



Of the species in the list the first three and the forty-sixth (G. rupestris) belong 

 to the flora of the Orient, the remainder of the north-west Frontier and almost all 

 the north-west Himalayan ones are species whose affinities are with the Altaian and 

 Siberian flora ; those of the Central and Eastern Himalaya, with very few excep- 

 tions, show on the other hand Chinese affinities. 



The number exhibited within brackets after the serial number of each species 

 in the list indicates to the student the serial number it bears in the Flora of British 

 India. 



1. (4.) Cortdalts diphylla. Wall. Tent. Flor. Nep. 54; leaves op- 

 posite long-petioled, twice ternately cut, primary petiolules not exceeding 

 the petiole in length; spur widely infundibuliform rapidly tapering, 

 obtuse slightly incurved at tip. Wall. Gat. 1430. C. longipes Don, 

 Prodr. 198 (not of DC.) C. Hamiltoniana Don, Si/st. Gard. i. 142. Cory- 

 dalis sp. Grif. Icon. PI. Asiat. t. 658, f. 3. C. rutaefolia H. f. fy T. 

 Flor. Lid, 262 ; Flor. Brit. Lid. i. 122 (not of Sibth. ) 



Central and Western Himalaya : Nepal, Wallich n. 1430 ! n. 1433 

 in part ! Kamaon ; common. Kashmir ; common. Hazara ; Stewart ! 

 Kurram" 'V alley, Duthie's Collectors ! Distrib. Afghanistan. 



This species has a globose tuber; the " long slender root" (De Candolle) or 

 " slender rootstock " {Hooker Sf Thomson) ascribed to it is in reality that part 

 of the stem between the deeply buried tuber and the surface of the soil. 



The species is easily distinguished from its nearest allies, G. rutaefolia, 0. Lede- 

 bouriana, G. persica, G. cyrtocentra, C. darwasica, G. macrocentra and 0. Sewerzovii 

 by its long-petioled leaves. 



Dr. Aitchison, in reaffirming M. Boissier's contention that the Afghan plant 

 united to this by Drs. Hooker and Thomson is different from G. rutaefolia, has not 

 called attention to the fact that, while this is the case, the Afghan plant to which 

 Boissier and he refer is even more distinct from the Himalayan one than it is 

 from true C. rutaefolia. In any case Dr. Aitchison's synonymy is slightly at 

 fault ; granting G. Grijjhthii to be the same as 0. rutaefolia H. f. & T. (not of 

 Sibth.), which is what he claims (Journ, Linn. Soc. xix. 151), the name G. diphylla 

 Wall., which is about 30 years prior to Boissier's, ought to have been used. Not 

 only however are the two species quite distinct, they are not even representa- 

 tive forms growing in distinct areas ; quite recently Mr. Duthie's collectors have 

 obtained true C. diphylla as well as G. persica (G. Griffithii) in the Kurram Valley, 

 while Genl. Gatacre on the other hand has collected G. persica in the Ziarat Valley. 



* Prantl and Kundig, in Engler, Naturlich. Pjlanzenfam. 



