14 THE SUBSECTION EU-CANINZ OF THE GENUS ROSA 
the styles, though they may be slightly combined, can be fairly 
easily separated, at any rate in the living plant. From the minor 
species of R. glauca those of R. canina can always be separated by 
the absence of a broad hemispherical woolly head of stigmas, and 
usually by the sepals and narrower stipules. 
. fi. dumetorum sp. agg. is the counterpart of R. canina, but 
with the leaflets hairy, even if slightly so on the midrib only. 
Rarely some varieties are slightly glandular on the secondary 
nerves ; these form the passage to R. Borrert. 
4. Ki. glauca sp. agg. has leaflets of medium or large size, never 
so small as in some of the varieties of R. canina. hey are 
always glabrous, and rarely glandular on secon nerves, and 
may be simply undly serrate. The stipules are broad 
and dilated upwards into large auricles. The peduncles are us y 
rise above the disc, and, even if only spreading, are much more 
persistent than in R. canina. The styles are very woolly, and 
the stigmas a a broad hemispherical mass, often almost as 
as the dise. a LORE TS 
R. coritfolia sp. agg. runs almost on { lines to R. 
5. R. parallel 
glauca, but its varieties are often smaller and more compact. The 
f. glauca, and the sepals more often reflexed. Some of the 
species are more or less glandular on the secondary nerves, such 
as . Bakert, or at least on the midrib, as in R. Watsoni and 
var. Lantont; but the general habit, rising sepals, and dense 
woolly head of styles seem to place them in this group rather 
than in RB. Borrert. 
My plan for dealing with the species and varieties in the 
British list is to quote first the author’s description, supplemented 
lise. It must, however, be borne 
t variations in the specimens of many of the 
notes thereon would occupy a volume, so I have almost co 
2 ee ee 
