Vol. Ill, No. 3.] On Gentiana coronata^ Rovle. 161 



li«t, have a lengtL of leaf that is four times the breadth, while 

 plants at the end of the list have leaves that are about twice as 

 long as broad. Between these two extremes lies every grade 

 that is intermediate. 



It may be remarked in passing that, despite Grisebach's in- 

 correct diagnosis (see Stapf in this Journal, 1906, p. 337), C. B. 

 Clarke intuitively placed Falconer's not-located specimen most 

 correctly. 



No. 5. Differing in the subacumiuate leaves, from the series 

 just defined, are specimens collected by C. B, Clarke and Dr, 

 Aitchison, The localities are :■ — 



Kashmir, south ofthe Val e. — Aliabad near the Pir 

 Panjal pass at 11,000 ft, ( C. B. Clarke, 28961 !). District of 

 R a w a 1 Find i. — Mari ( Hb. Aitchison !), 



^ h ^ 



The flowers of both are fimbriate, and relatively large : 

 the capsule is conspicuously crested. Unfortunately only one 

 of the eight plants that lie to* my hand has root-leaves. This 

 one, collected by Mr, Clarke, I figure. 



No. 5a, Next I refer to an erect, firm-leaved, unbranched plant 

 collected in : — 



r 



Kashmir beyond the Indus, — Gilgit (Tanner, 112rt!) 

 Sui in Gilgit (Tanner, 112 !). 



It connects G, marginata with 6f. marginta, van recurvatuy 

 having leaves somewhat like the latter, but the habit of the former. 

 The typical lower leaves are 10x5 mm. and the margins are 

 cartilaginous. There are no fimbriae in the flowers. 



We now leave one of the series diverging from Hooker^sand 

 Thomson's G. marginata for the other. The reader may think of 

 them like this 



TYPE SA , ' 



HUCEU^ ' TYPE b TYPE 6 



TVPt 7 



TYPt 7a 



TYPE 9 



RECUnVATA 



No. 6\ Conspicuously herbaceous is the next plant, and with 

 the scarious margin of the sepals reduced to a very narrow line. 

 It branches from the root, and has lanceolate or lanceolate-orate 

 leaves. Its localities are : 



