38 



A LIST OF BRITISH ROSES 



Leaflets Uniserrate. 



E. OBTUSIFOLIA Desv. Journ. de Bot. ii. p. 317. I think, after 

 seeing growing bushes of this, which I had not seen when I wrote 

 my paper on the Eu-canincB, that it must certainly go to the 

 present group. To so place it would be quite in accordance with 

 my views that species which look alike should be associated, 

 though their technical points may differ. The resemblance is so 

 great that I cannot distinguish B. ohtusifoUa in the field from 

 B. tomentella without examining the leaf serration. My former 

 views as to its classification have been strengthened by what I now 

 believe to be the misnaming of many examples of the Dumetorum 

 group by Sudre and by British authors. It is a frequent species, 

 at least in Surrey, and I have little doubt that its distribution is 

 general. B. canina var. frondosa Baker is synonymous. V.-c. 2, 

 3?, 4?, 17, 34, 36, 40?, 58?. 



E. CANINA var. concinna Baker, Monogr. p. 238. Since writ- 

 ing my former paper (E. p. 83) I have seen Baker's type in 

 Borrer's herbarium. It is from South Devon, and is in all 

 respects like B. ohtusifoUa, but with hispid peduncles. I have 

 not seen it from elsewhere. V.-c. 3. 



SUBSECTION VILLOS^. 



This subsection is our most difficult one, not so much in segre- 

 gating the species, which is difficult enough even when the group 

 is known, but in assigning a very large proportion of herbarium 

 specimens to their proper group. In the growing plant the habit 

 may be useful, but in most cases the behaviour of the sepals on 

 the ripe fruit is indispensable. In addition to this fundamental 

 uncertainty, the aggregate nature of some of our old species, such 

 as B. mollis Sm., B. Sherardi Dav., B. tomentosa Sm., &c., is a 

 source of trouble ; moreover, the British treatment of the sub- 

 section is on different lines to that followed on the Continent, so 

 that the names cannot be correlated. It would appear, also, that 

 the subfoliar glandular development is of much less importance in 

 this than in the Eii-canincB. I scarcely know any species in 

 which examples cannot be found which are similar to one anotlier 

 in all respects, but some have a full and conspicuous clothing of 

 subfoliar glands, which in others is almost or quite absent. 

 Much the same remarks, though to a less degree, apply to the 

 curvature of the prickles and shape of the fruit. 



I am greatly indebted to Mr. Barclay and to the late Mr. Ley 

 for a considerable supply of specimens and notes, without which 

 my knowledge of the subsection would have been indeed meagre. 



GROUP POMIFERA. 



SUBGROUP POMIFERiE. 



If this subgroup can be separated from that of Molles at all, 



it is by its thin, large leaflets, but typical B. pomifera has, on the 



Continent, a whole series of varieties, which closely connect the 



two subgroups. 



