Hybridism and Classification in the Genus Rosa . 157 
difficult to diagnose as such, since the supposed parents are so 
closely similar in many features. There is nothing improbable, 
however, in the supposition that closely allied species may hybridise 
just as freely as the species mentioned in the foregoing paragraphs. 
This theoretical consideration is of great importance and it bears 
very directly on the whole discussion which follows. If closely 
related species or species and varieties hybridise the hybrid origin 
of the offspring would be exceedingly difficult to detect. In actual 
practice they would not, as a rule, be distinguished as hybrids, but 
would be considered distinct species or varieties and would be named 
as such. Thus, perhaps, has arisen a large portion of the total 
number of named varieties of roses. It is exceedingly interesting to 
find, therefore, that Almquist, one of the most recent writers on the 
genus, treats R. glauca and R. coriifolia , two very closely related 
forms, as a collective species under the name R. Afzeliana Fr.= 
R. glauca X coriifolia (see Traaen, 1913). 
On the general question of hybridism in the Rosaceae the work 
of Jeffrey may be mentioned. “ Not only are certain of the 
Rosaceae recognisable as hybrids on account of their transitional 
features of organisation, Mendelian phenomena, etc., but certain 
others which have not revealed themselves as hybrids in these ways 
are clearly such as a result of a study of their spores. We have 
consequently to distinguish in the case of the Rosaceae, if the 
morphological features are taken into account as well as the data of 
systematic botany, three kinds of individuals, namely, pure species, 
recognised species which from the condition of their pollen are in 
reality concealed hybrids or crypthybrids, and recognised hybrids.” 
The subject, in so far as it concerns the genus Rosa, has been 
pursued by Miss Cole (1917) who, from a study of the pollen in 
numerous roses, concludes that the genus contains only a few pure 
species, while the great majority of so-called species are really of 
hybrid origin. Unfortunately, Miss Cole’s investigation does not 
include many species native in Britain, and the condition of the 
pollen in British forms I hope to make the subject of a future 
paper. 
It has already been mentioned that there is no experimental 
evidence to show whether the species-hybrids described above 
segregate or remain stable. It has generally been believed that 
only crosses between varieties segregate, but the work of 
Brainerd (1907) on American Violets points clearly to segregation 
in the offspring of species-hybrids. Brainerd found growing wild 
