CLASSIFICATION -- 13 
The species I have adopted are very nearly those of Keller, 
viz. . Borreri Woods, R. canina Linn., R. dwmetorum Thuill., 
Vill., and R. coriifolia Fr. Keller gives very long and 
group 
tead of being of the aggregate species so as to exclude the 
ups. 
of the aggregate species under which I group the 
British varieties are as follows. Though I do not regard their 
subordinate species, nor their varieties, as being respectively of 
equal rank, I have not attempted any division into subspecies, 
_1. &. Borrert sp. agg. I regard as including all plants with 
more or less strongly biserrate leaflets, almost always rather or 
almost eglandular bene ese two might equally well be 
c er R. dume . 
2. RB. cani a is includes all species with leaves quite 
: na sp. agg. Thi 
glabrous on both sides, though they may be glandular on the 
Secondary nerves, as in R. scabrata Crép. The leaflets may be 
e 
S are usually hispid or glabrous, rarely woolly, and are 
ed in a subcylindrical column, rarely projecting much, and 
usually not at all, from the disc. Individuals with subeolumnar 
