16 THE GENUS FUMARIA IN BRITAIN 



taken from fresh specimens cultivated in the Botanical Garden at 

 Lund. Hammar places F. Gussonei as a species among his 

 CaioreolatcB, where F. confusa also stands, associated with F. 

 BorcEi and F. muralis as varieties of F. media Lois. On the 

 other hand, F. affinis and F. vagans are made distinct species in 

 the section Agrarice. With Haussknecht, I quite fail to under- 

 stand why Hammar should join the rugose-fruited F. confusa 

 with F. Bormi and F. muralis, but the other discrepancies be- 

 tween the two monographers seem less inexplicable. Hammar 

 placed his F. affinis and F. vagans among the AgraricB owing to 

 the form of the lower petal, which may easily be mistaken in 

 dried specimens, and becomes essentially obscure in ill-grown 

 plants. There can be little doubt, judging from his measure- 

 ments, but that most of the specimens in this group examined by 

 Haussknecht showed poor flowers, and with these in the dry 

 state he probably did not suspect that any of them could produce 

 a lower petal of the Agrarice type, although such may be found, 

 under favourable conditions, not only in F. affinis and F. vagans, 

 w^here it was observed by Hammar, but also. in the British i^. 

 confusa, and even in typical F. Gussonei. 



In this respect then, as well as in the rugosity of the fruit, 

 these forms united by Haussknecht under F. Gussonei tend to 

 approach the subsection Agrarice, to which a further resemblance 

 is seen in some cases in the absence of the dark-coloured tip of 

 the upper petal. 



Of the smooth-fruited species, F. Beuteri shows an unmistak- 

 able likeness to F. Thureti Boiss. and F. anatolica Boiss. of the 

 subsection Latisejjalce of the Parviflorce, and to this F. Munbyi 

 also is probably less closely allied, although it is regarded by 

 Hammar as synonymous with F. Borcei. The remaining species 

 show more affinity with the Capreolatcs. The relations of F. 

 muralis and F. Borcei, reduced to varieties of F. media Lois, by 

 Hammar, were dealt with at some length in my paper in Journ. 

 Bot. 1902, and now require a further elaboration owing to the 

 great diversity of forms occurring in these islands. F. sepium, 

 which is kept up as a species by both monographers but in 

 Eouy & Foucaud's Flore de France is made a synonym of F. 

 muralis var. ijlatycarpa, is a rare Spanish plant, chiefly remark- 

 able for its lanceolate sepals, of which I have seen no authentic 

 specimen. 



'■^'- Fruit, when dry, smooth or nearly so, rarely distinctly 

 rugulose at maturity. 



3. FuMARiA MURALIS Sonder. 

 The determination of the plant found by Sonder near Ham- 

 burg and described under this name w^as dealt with at length in 

 Journ. Bot. vol. xl. pp. 175 sqq., where it was shown to differ from 

 similar slender forms growing in Britain that had formerly been 

 referred to the same species. These slender forms were placed 

 as varieties under F. Borcei Jord., which, owing to the large series 

 of intermediates, was itself made a subspecies of F. muralis, the 



