SUBSECTION VILLOS 53 
ate of very ee eeteons material under one name, or the 
reation of new species or varieties, or at least the adoption of 
Cand canta! names, caw unknown, or little known, to Britis 
botanists. 
Following my plan ee for ae ee of these papers, I 
Saat as their authors named them, but I have ranged them 
sential the sronge to which they must, in ee opinion, eventually 
e refe 
I aera for the purposes of this paper and not as a final 
arrangement, to subdivide the subsection into three groups, viz. 
f. pomifera (including R. mollis), R. Sherardi and R. tomentosa. 
Most authors make two su ubsections, one containing R. ponies 
and f. mollis, the other f. omissa (which includes RB. S dt) 
and &. tomentosa. The main subsectional difference relied tae 
by such authors is the relative persistence of the sepals, combined 
with other and less stable features, which on the whole appear 
to be - less importance than oe “differences — two = 
groups as FR. canina and R. coriifolia. It s hardly ¢ 
sistent, therefore, to sabiunds the Villose coke ene aioe 
an groups, especially as that of Sherardi is exactly intermediate 
beta the other two, and all shade off imperceptibly into one 
anot. 
Bry the groups may be defined as follows :— 
of #. ifera. Of low growth. Stems rigid, with 
incurved auricles (acco me to Conkihental authors). Peduncles 
usually short. Flowers deep rose. Fruit we soft an 
pulpy early, crowned by the subentire persistent sepals, which 
adhere till the fruit itself decays or falls off. Styles villous. 
* The prickles on the middle —_ of the hoa onthe Howaning hoot then 
