A LIST OP BRITISH ROSES 33 



globose fruit and leaflets always glandular beneath, and from the 

 latter in its more hairy, less broadly oval leaflets. Besides those 

 mentioned in E. pp. 103, 104, I have only seen two specimens, 

 both from Perth {Barclay). The sepals are considerably spread- 

 ing in both, and one has its primordial fruits turbinate, but I 

 think both are correctly referred here. A specimen from N. E. 

 Yorks {Groves, Bot. Exch. Club, 1906) was labelled B. coriifolia 

 var. ostensa Gren. by Sudre. It seems nearest var. Lintoni. I 

 have in my herbarium a specimen from Braemar {E. F. Linton), 

 in which the hairs on the leaflets are almost confined to the mid- 

 ribs beneath, and its sepals are loosely reflexed on the young fruit. 

 It agrees in other respects with the type, and is probably correctly 

 named, but had I seen it alone I should certainly have referred it 

 to the next subgroup. V.-c. Perth, 92. 



R. CELERATA Baker, Rev. p. 31. If we confine this species to 

 examples witli leaflets like those of B. tomentella, i. e., rather 

 small, broadly oval, or suborbicular, fully biserrate, and glandular 

 on the primary veins beneath, subglobose or broadly ovoid fruit, 

 with suberect sepals and a broad, woolly head of styles, it should 

 be fairly easily recognized, but I have seen no specimens except 

 those mentioned in E. p. 103. V.-c. 39. 



R. Watsoni Baker, Rev. p. 29. This is best recognized by its 

 leaflets being rather large, broadly oval, and eglandular beneath, 

 its fruit ovoid or subglobose, and its sepals suberect. It is near 

 B. celerata, but besides the difference in leaflets, it has much less 

 glandular petioles. The Perth specimen {Barclay) is off type in 

 having somewhat glandular peduncles, and leaflets somewhat 

 narrowed below, which two characters take it towards B. Bakeri. 

 V.-c. 65 ?, 69 ?, 70, 88. 



SUBGROUP SUBCOLLIN^. 



This subgroup contains species with reflexed or spreading 

 sepals, and like the corresponding subgroup of the group Glauca, 

 it is, I think, more extensive than has been supposed. As with 

 all other characteristics, it is difficult to draw the line between 

 reflexed, spreading and suberect sepals, but it is obvious that it can- 

 not be drawn at the same place for each species, as in that case 

 specimens identical in all other respects would be separated. 



Leaflets Unisbrrate or nearly so. 

 R. c^siA Smith, E. B. tab. 2367. In my Eii-canincB I followed 

 D6s6glise and Keller in placing this species in the Dtimetoruvi 

 group, on account of its reflexed sepals and not very woolly styles, 

 but after correspondence with Mr. Barclay, I agree that this is its 

 best place. I have seen no further specimens beyond those men- 

 tioned in E. p. 87. Specimens from Dovedale, Derbyshire, have 

 been i-eferred here by various botanists. I have one of these forms, 

 which differs from the type in its much larger leaflets and its 

 decidedly glandular peduncles and even fruit. It is a poor example 

 to found a record upon, and looks very different from the type, 

 Journal of Botany, Nov. 1911. [Supplement.] d 



