216 ENGLISH BOTANY. 



or less pinnatifid, generally glabrous, but sometimes with gland- 

 tipped setae on the outside.* 



Mr. Baker distinguishes the following forms, which he arranges 

 under four groups. 



Group I.— SUBRUBIGINOS.E. 



Leaves sparingly glandular beneath, either all over or only on 

 the principal veins. Pedicels naked or with aciculi. Fruit various. 



The plants of this group are placed by M. Deseglise in the 

 section Rubiginosse. They appear to me to be best placed with 

 E-. inodora and Borreri, except E.. vinacea, which appears to be a 

 true Dog-rose. 



1. Rosa Blondseana. Ripmrt. 



Baker, in Nat. 1864, p. 103. 



Stem with divaricate branches ; prickles moderate, slightly 

 curved. Leaflets elliptical or obovate-elliptical, glaucous above, 

 glabrous and glaucous with scattered glands (which are most 

 abundant on the midrib beneath), sharply doubly serrate, with the 

 secondary serratures gland-tipped. Petioles glandular, destitute of 

 hairs. Stipules and bracts glabrous, ciliated with gland-tipped 

 setae. Pedicels with a few aciculi and gland-tipped setae. Styles 

 thickly hairy. Pruit naked, obovate or subglobose, turning 

 scarlet in September. Sepals leaf- pointed, pinnate, with gland- 

 tipped setae on the outside, some of them falling by the time the 

 fruit changes colour. 



In hedges. Kilvington, North-east Yorkshire ; and Mr. Baker 

 has gathered a very similar plant in the counties of Perth and 

 Aberdeen. 



* I am indebted to Mr. Baker for specimens of the greater number of the forms 

 of the Dog-roses which he has so carefully worked out and compared with Continental 

 specimens named by M. Deseglise. I however feel myself quite unable to come to any 

 satisfactory conclusion as to how many of the forms ought to be regarded as subspecies ; 

 indeed this, as in the case of the Rubi, can only be ascertained by continued cultivacioa 

 from seed, so as to ascertain what forms are really permanent, as it is impossible to 

 determine h priori what characters invariably descend unchanged from the parent to 

 the olTspriug, 



