THE GENUS FUMAEIA IN BRITAIN 3 



far as I can ascertain, exclusive of forms either recognized as 

 hybrids or subsequently reduced to the rank of varieties. Of 

 these, one is from Morocco, one from Spain, one from Italy, three 

 from Sicily, one from Transsilvania, and two (my F. purpurea and 

 F. occidentalis) from Britain. 



It is noteworthy that in a monograph of the Italian Ficmari- 

 acecB, pubhshed at Florence in 1897 by L. Nicotra, the author has 

 reverted to Hammar's subdivisions of the genus, but with the 

 section Officinales placed after the Gapreolatce and Agrarice. 



It will thus be seen that Fumaria must be regarded as a 

 polymorphic and critical genus of plants, concerning which mono- 

 graphers have shown considerable differences of opinion in the 

 grouping of the species. Indeed, as in some other critical genera, 

 it seems more difficult to define satisfactorily the groups of forms 

 than to describe these individual forms as species irrespective of 

 their allies. 



The best and most natural classification on the whole, although 

 it is not free from objections, seems to be that of Haussknecht, 

 in adopting which I have enlarged his definitions in the hope of 

 making them more exact ; and, as his sequence of species has 

 not been generally followed in recent Floras, I place the Latisectca 

 before the AngustisectcB, and begin with the Gapreolatce, after the 

 example of Nicotra's monograph, and Eouy & Foucaud's Flore 

 cle France (vol. i. p. 171). This order has the further incidental 

 advantage of necessitating the least change from that of our 

 present British text-books. 



It may be well to recur briefly to the necessity of exercising 

 judgement in the collection and examination of specimens. With 

 nearly all the species the habit and foliage vary exceedingly 

 according to surrounding conditions — so much so that, although 

 they are generally distinctive, I cannot always readily recognize 

 as specific characters the features attributed to them in some of 

 the Continental Floras. It is of the first importance throughout 

 the genus that normal flowers, fully coloured and with the corolla- 

 wings developed, should be obtained. It is a remarkable charac- 

 teristic of the Fumitories — especially the LatisectcB — that when 

 they grow under unnatural or starved conditions the corolla fails 

 to develop and tends to revert to a small primordial form, whitish 

 in colour, except for the green keels of the two outer petals, and 

 with the wings of these two petals obsolete. Such corollas are 

 useless for the determination of species. A corresponding de- 

 generation, however, does not take place in the calyx, which, 

 while variable within certain limits, I believe nearly always main- 

 tains its essential characters ; and the same may be said of the 

 fruits, except that sometimes, in the case of weak or exhausted 

 plants, they may not reach their ordinary dimensions. The form 

 of the pedicels and their bracts is also important, as well as the 

 length of the peduncles bearing the racemes of flowers. A 

 curious feature of the corolla in Fumaria is the dispersion or 

 " running " of the purple colour of the apex after fertilization and 

 as the flower fades. 



