164 
J. R. Matthews. 
We may pass now to the glanca-coriifolia groups and 
proceed to an analysis precisely similar to that which has been 
applied to the canina-dumetorum series. If we take first the 
eight combinations which include h, viz. hBGS, hBGs, hBgS, 
hBgs, hbGS, hbGs, hbgS and hbgs, we shall be dealing with 
glauca forms. The combination hbgs is R. Reuteri God. and the 
hispid-peduncled form hbgS is R. Reuteri var. transiens Gren. 
R. subcristata Bak. is the form hBgs and R. fugax Gren. corres¬ 
ponds to hBgS. The biserrate forms with subfoliar glands are 
hBGs and hBGS, the former combination having been named R. 
stephanocarpa D^segl. et Rip., the latter being R. glauca var cenensis 
Kell. Only two of the eight combinations remain unrecognised, 
namely hbGS and hbGs. Turning now to the eight combinations 
having H in their composition, viz. HBGS, HBGs, HBgS, HBgs, 
HbGS, HbGs, HbgS and Hbgs, we find that six have been named. 
These, of course, come under the coriifolia group. R. incana 
Kit. is an Hbgs form and constituted the type of Fries. The 
combination HbgS is R. bovernieriana Lagg. et Pug. R. Watsoni 
Bak. and R. caesia Sm. correspond to the forms HBgs and HBgS 
respectively. The combination HBGs applies to R. coriifolia var. 
Lintoni Scheutz and the counterpart with glandular peduncles, 
HBGS, is R. Bakeri Desegl. Thus, in the glauca-coriifolia 
groups, twelve of the sixteen theoretical combinations have been 
discovered and have received distinctive names. Unfortunately, an 
analysis on the basis of glaucousness and non-glaucousness or 
greenness cannot be attempted since the distribution of these 
characters in all the forms of the two groups is not known. But 
these characters are unquestionably important ones, and they 
receive much emphasis in Almquist’s classification of Swedish 
roses to which reference will be made later. 
The names R. subcanina and/?, subcollina date from 1873 when 
they were used by Christ in Rosen der Schweiz , the former to describe 
a variety of R. Reuteri (glauca), the latter as a variety of R. coriifolia. 
By some authors they are still regarded as varieties or sub-species, 
but there are certain advantages in treating them as independent 
groups equal in value, from the point of view of the systematist, to 
R. glauca and to R. coriifolia. Already, evidence is accumulating 
to show that they probably each comprise a series of forms 
developing along lines parallel to those dealt with for the two 
aggregate species of which they were respectively described as 
varieties. The type of R. subcanina (Chr.) is a uniserrate form, but 
