18 THE GENUS FUMARIA IN BRITAIN 



Some of these fumitories are highly critical, and differences of 

 opinion may arise as to their exact affinities. In the British 

 Islands they have reached a very varied development, which 

 perhaps obtains in no other country and was never suspected by 

 Haussknecht when, without any hesitation, he ranked F. muralis 

 and F. Borcei as distinct species in his monograph. Some of the 

 variations are w^ell marked, and readily admit of precise defini- 

 tion, but gradations in habit, flower and fruit are so numerous and 

 irregular that a clear line of demarcation between F. muralis and 

 F. Borcei, even as subspecies, is by no means easily drawn. The 

 difficulty is increased by the recognition of the globose-fruited 

 plant of Madeira as a variety of F. muralis, as this destroys the 

 value of the pointed fruit as a clear subspecific character. But 

 the only view^ that can reasonably be held, when it is remembered 

 that the Atlantic Islands, and not North Germany, are the head- 

 quarters of the species, is that the Hamburg type is a local form 

 of a plant subject to considerable variation. On the other hand, 

 as F. BorcEi (of which nearly all the foreign material both at the 

 British Museum and at Kew is French) is a plant of Western 

 Europe, and is perhaps nowdiere so abundant as in Britain, the 

 variants that occur here, though often more or less tending 

 towards the type of F. miiralis, must generally be regarded as 

 local modifications of F. Borcei, and I therefore adhere to the 

 retention of these forms under this latter subspecies, as shown in 

 my paper of 1902, albeit some of them superficially bear the 

 facies of F. muralis Sonder. 



So far as I am aware, the only named varieties of F. Borcei 

 published prior to 1902 are those in Clavaud's Flore de la Gironde 

 (p. 47), which I adopted in my earlier paper. These varieties are 

 described in some detail, although little is stated respecting 

 F. Borcei itself, and their alleged deviations from Jordan's type 

 correspond closely with variations to be seen in Britain. As 

 Clavaud's figure of his variety muraliformis, however, recalls the 

 globose- fruited plant of Madeira, and as F. muralis is recorded for 

 the Gironde district by Haussknecht, as well as in Eouy and 

 Foucaud's Flore de France, I have obtained for inspection, 

 through the kindness of M. Beille, Clavaud's actual specimens 

 from the Jardin des Plantes at Bordeaux. These specimens 

 represent the varieties muraliformis and verna only, var. serotina 

 being absent from Clavaud's herbarium. Of the variety murali- 

 formis there are two small plants, one fairly normal and showing 

 both flowers and fruit. This resembles the British plant to 

 which I have applied Clavaud's name, except that its flowers are 

 smaller and the fruits, though somewhat obovate, no larger than 

 in typical F. muralis. The other variety, verna, is represented by 

 two small and young plants obtained early in the year from 

 fallows. Only one of them shows any fruits, and these, which 

 are possibly not quite mature, are of muralis size, globular and 

 apiculate. The flowers, which appear normal, are in one example 

 11 mm. long, and in the other 2 mm. less — in both smaller than 

 an average good flower of typical F. Borcei. As Clavaud states 



