22 A LIST OP BRITISH ROSES 



in some bushes it is difficult to find any otlier characteristic, which 

 gives considerable weight to the idea that many of them may be 

 mere varieties of that group. Many of the species of the present 

 group are mainly distinguished from one another by increasing 

 development of hair on the leaflets ; other written characters 

 exist, but they are so variable that they are difficult of application. 



SUBGROUP EU-DUMETORUM. 



Simply or irregularly serrate leaflets and smooth peduncles are 

 the distinguishing features of this subgroup. If attention be 

 paid to the presence of hairs on the under surface or midribs of 

 the leaflets, not on the petioles only, such as are frequently seen 

 in the Canina group, there should be no difficulty in recognizing 

 its members as a whole, but their separation into species or 

 varieties is difficult. 



E. DUMETORUM ThuiU. Fl. Par. p. 250. This species should 

 have its leaflets hairy on both sides, even when old, and sub- 

 globose fruit. It is very difficult to separate from B. urhica, and 

 often, at least in the herbarium, from B. obtusifolia. Most con- 

 tinental authors distinguish it from the former by little else than 

 its more hairy leaflets, with unarmed petioles and less hairy 

 styles, while it is marked off from B. obtusifolia, when the group 

 characters of the latter are not well-marked, chiefly by its rose- 

 coloured flowers. Yet Thuillier's description credits it with 

 " petioles minutely prickly all round," and "flowers rose or white." 

 Keller treats it as a group name only, a treatment which Crepin 

 had already suggested. The vice-county distribution here given 

 is chiefly that of forms which cannot be definitely referred to any 

 other segregate. V.-c. 2 ?, 3, 32, 36, 58, 88. 



E. suBMiTis Gren. in Schultz Arch. p. 332. This is the 

 common form of B. dumetorum (agg.), of which it may be regarded 

 as an ovoid-fruited variety, but although Thuillier described the 

 fruit of his species as spherical, the only example I have seen of 

 his gathering has it ovoid or ellipsoid. Long peduncles and very 

 spreading or even decidedly rising sepals, often rather long per- 

 sistent, are frequent in specimens referred to B. suhmitis, and 

 some of them have slightly biserrate leaflets. The rising sepals 

 cause these to be sometimes mistaken for members of the Corii- 

 folia group, but their thinly hispid styles and long peduncles 

 should indicate the group. Two specimens "of this form, from 

 Eadnor, with decidedly erectish sepals, have been seen by Crepin, 

 who considers them to belong to the present, and not to the 

 Coriifolia group. On the other hand, both Sudre and Dingier 

 consider a Cheshire form, which is practically identical with that 

 from Eadnor in peduncles, fruit, sepals, and styles to belong to 

 the Coriifolia group. I prefer Cr6pin's opinion. V.-c. 2, 17, 36, 

 43, 58, 62. 



E. Gabrielis F. Ger. ex Magn. Serin. Fl. iv. p. 84. I have 

 not been able to gain access to Magnier's work cited, nor have I 

 seen any continental examples, so can only form an idea of it 



