THE 
New Phytologist 
Vol. XIX, Nos. 7 & 8 . July & Oct., 1920 . 
[Published August 24th, 1920.] 
HYBRIDISM AND CLASSIFICATION IN THE 
GENUS ROSA. 
By J. R. Matthews, M.A., F.L.S. 
Lecturer in Botany , Birkbeck College , London. 
I T is generally recognised by students of the British Flora that 
the genus Rosa presents interminable difficulties in connection 
with the classification of its innumerable forms into anything like a 
satisfactory or natural system. This fact, it seems to me, is 
reflected in the ever-increasing number of names which go to swell 
an already lengthy British Rose List. Scarcely a year passes which 
does not witness the publication of new “species” or “varieties” 
which are considered by their authors as deserving distinctive names 
because the plants show a few differences, very often only minor 
ones, from allied species. 
Among the earlier workers we find that Baker (1869) in his 
Monograph of British Roses is content with thirteen species (eleven, 
if we omit R. rubella Sm. and R. pomifera Herrm. which are 
doubtfully native) and a moderate number of varieties. On the other 
hand, the London Catalogue of British Plants (1908) enumerates 
twenty-five native species and a large number of varieties, many of 
the additional “ species,” of course, being Baker’s varieties, while in 
the List of British Roses by Wolley-Dod (1911) about one hundred 
and seventy names appear. Since this author expresses no opinion 
as to the relative value of the names, it is difficult to say how many 
are intended to stand as species. 
There are probably no plants more variable than roses and it 
is perhaps not too much to say that no two bushes of the same 
“ species ” are quite alike in all the technical characters relied upon 
by rhodologists for making a diagnosis. In the case of many micro¬ 
species it is found on investigation that authentic specimens collected 
or named by the authors themselves neither agree with the original 
