12-t TRANSACTIONS OF THE [Sess. lxxxvi 



When Franchet wrote the above he had no doubt in 

 mind the Lilies of Japan and the maritime provinces of 

 China, but from the explorations of Wilson and of Forrest 

 we know that some of the indigenous species of Western 

 China have been cultivated for many generations as an 

 article of food Cultivation in such extensive provinces 

 as Szechwan and Yunnan with their great divergences in 

 climate has no doubt tended to produce cultural races 

 in the species used for food. Franchet's caveat must 

 therefore be borne in mind in all questions dealing with 

 the delimitation of the species of the area. 



My second extract from Franchet is as follows : — 



" Mais comme, en meme temps, tons les groupements des 

 especes de Lis ont pour base cette forme du perianthe, il 

 en resulte de serieuses difficultes lorsqu'il faut comparer 

 une forme cultivee avec son type spontane, ou ce que Ton 

 considere comme tel. La difficulte est encore augmentee 

 par ce fait que la direction dressee ou revolutee des divisions 

 du perianthe ne se manifeste completement que tardive- 

 ment et, en general, seulement apres le fecondation." 



Baker's division of the known Lilies into five groups 

 (Journ, Linn., Soc. xiv (1875), p. 225) affords a useful con- 

 spectus of the species, but all who have dealt with the 

 genus know that there are several species which seem to 

 hover between two groups and even form a transition 

 from one to the other. I think it therefore apposite to 

 recall Franchet's words. 



Another element of complexity is introduced by the 

 peculiar geographical conditions of Western Yunnan and 

 to a certain extent of Szechwan. Recent work on the 

 flora of these regions tends to show that there are several 

 quite distinct geographical areas in these regions and that 

 what at first appear to be conspecific plants in adjoining 

 areas show, after investigation, difierences which are not 

 far from specific. Yunnan, owing to the peculiarities of 

 its natural features, appears to be rich in geographical 

 races. This separation of areas seems to me to be due in 

 great part to the depth of the valley floor separating one 

 mountain range from another. In a great number of these 

 Yunnan vallevs the floor descends to an elevation little 



