160 
J. R. Matthews. 
may possibly indicate that R. pimpinellifolia may have entered 
into the plant’s composition through hybridisation. By some 
authors R. gallicoides is regarded as R. arvensis x gallica. 
Rosa stylosa Desv. in Journ. de Bot. 1809, is the only aggregate 
species of the Stylosae section, a group which is unquestionably 
intermediate in its general features between the Synstylse and the 
Eu-caninae, some authors including it in the former section while 
others place it in the latter. This intermediate position 
immediately suggests hybridisation between Synstylae and 
Eu-caninae forms in the production of those forms which constitute 
the Stylosae. The distribution of R. stylosa agg. is the same as 
that of R. arvensis agg., while Eu-caninae forms of the group 
dumetorum are widespread and it is from this group of the 
Eu-caninae that some members of R. stylosa are almost 
indistinguishable. It might be interesting and instructive, 
therefore, to cross systematically R. arvensis and R. dumetorum. 
In R. stylosa there is little variation in the serration of the 
leaflets, all the British forms being uniserrate except var. evanida 
Chr. which is biserrate. But the leaflets in some are hairy (H), in 
others glabrous (h). Forms with rose-coloured (R), others with 
white (r) flowers occur, while some have setose (S), others smooth 
(s) peduncles. If we deal with those three characters only, the 
following eight combinations would be visibly different—HRS, 
HrS, HRs, Hrs, hRS, hrS, hRs and hrs. At least five of these 
combinations have been met with in Britain and are known by 
distinctive names. The type of R. stylosa Desv. is the form HrS, 
while HRS corresponds to R. systyla Bast. Hrs is R. stylosa var. 
corymbosa Desv. = var. opaca Bak. and var. pseudo-rusticana Crep. 
is the combination hrS. R. virginea Rip. comes under the hrs 
combination. Only one form with rose-coloured flowers is included 
in this analysis, but another, viz. R. parvula Sauz. et Maill. occurs 
on the Continent and has been reported as occurring in Britain. It 
corresponds to the combination HRs. 
We pass now to Section III which is a very large one, the 
subsection Eu-caninae itself containing by far the largest number 
of our British species and varieties. In general, the members 
of this subsection can be distinguished by their large size and 
arching habit, their uniform and decidedly hooked prickles and by 
the styles being free, at least not agglutinated into a column like 
those of Synstylas and Stylosae. Since the division into groups 
given in Table 1 is not that most usually adopted by systematists 
