252 
27. R0SACEJE. 
Tribe III. Rosece. 
8. Rosa L. 
Rose. 
§ Leaves and shoots without glands or seta, 
a. Styles distinct. 
1. R. canina L. Rosa brava. Dog Rose. 
Branches naked smooth ; prickles uniform strong hooked 
compressed remote or scattered; lfts. 5, 7, 9, subcoriaceous, 
mostly smooth and naked without glands, simply or doubly ser- 
rate ; stip. adnate persistent ; fl. aggregate ; sep. pinnatitid de- 
ciduous. — DC. ii. 613; Koch 250; Bab. 112; Brot. ii. 340 (in 
part) ; Lindl.Ros. Mon. 98; Hook. Fl. Sc. i. 157 (in part); Hook. 
Br. Fl. (ed. 5) 106 ; Sm. E. Fl. ii. 394 ; WB. ii. 16. 
a. glabra Desv. ; petioles with a few scattered prickles, nearly 
or quite smooth ; lfts. naked and quite smooth, keeled, simply 
serrate, elliptic-ovate or oval, more or less pointed at each end ; 
pedic. and ovate fr. quite smooth. — DC. 1. c. R. canina a, 
Lindl. Ros. 1. c. ; Hook. Br. Fl. 1. c. ; Bab. 1. c. R. canina a. 
vulgaris WB. 1. c. R. canina EB. t. 992. — Shr. per. Mad. reg. 
2 (upper part) and 3, f. Ravine W. of the Allegria near 
Funchal ; ravines between Camacha and S. Antonio da Serra ; 
descent from the Mount into the Rib. de S ta Luzia ; about the 
Primeira Vista do Curral and thejJardim da Serra ; plentifully in 
the upper part of S. Vicente and of the ascent to the Paul da 
Serra below the Tanquinhas; Rib. Frio, da Metade, &c. June, 
.1 illy — Almost precisely the genuine common English Dog Rose 
with more orless pale-pink- or blush-coloured fl., in Mad. always 
nearly white, with the faintest tinge possible of blush, and fra- 
grant. Upper or flowering branches only here and there or not 
at all pricidy. lifts, mostly shorter broader and rounder or less 
elliptic than in the English pi., being rarely twice as long as 
broad, often much less, and scarcely or not at all pointed at the 
base ; very smooth and shining, full gr., a little paler beneath, 
notin the least glaucous or pubescent ; their serratures simple 
erect, subconnivent upwards. Styles short, distinct to their 
very base, sometimes quite smooth, sometimes villose-hairy, 
and thus not confirming a remark of Dr. Lindley (Rosar. Mo- 
nogr. 103) quoted in Hook. Fl. Scot. 15&, to the effect that, 
tin* further to the south the pi. is found, the smoother are its 
styles, and that they are entirely without hair in Madeira. 
So in another var. (R- canina y. Armida WB. ii. 16, which is 
very nearly if not quite identical with R. sarincntacea Woods, 
KBS. t. 2595 = R. canina /3, Hook., Bab., &c.) which I found 
