Rose Forms, as determined by their Cytological Behaviour . i 7 1 
acting together may take part of the cytoplasm to themselves, to develop 
in the end into pollen grains of dimensions varying with the chromatin 
content of the nuclei, whence they originate. 
After this, in the bulk of the pollen grains, degeneration sets in, 
although a fair proportion, as proved by direct observation, are quite 
functional. 
Giving due weight to the exceedingly close resemblance between the 
whole of the events during microspore formation in R. coriifolia and 
R. Sdbini , from the typical preparatory stages to the anomalous later ones, 
and recollecting the perfectly ordinary and orderly meiotic phase in 
R. arvensis , we are forced to the opinion that the same agency, hybridity, 
is responsible in the first-named pair of plants for their common peculiarities. 
Undoubtedly, the very least one can say is that the cytological behaviour 
of R . coriifolia is that of a hybrid. 
(3) Rosa coriifolia var. Lintoni , Scheutz. 
This rose is one of the glandular members of the Afzelianae and 
would, by Almqvist, be regarded as appertaining to the Rujfiginosae. We 
prefer, for reasons perfectly apparent when the rose is studied as it grows, 
to leave the rose as placed by English rose students. 
Its somatic chromosome number, as in the case of the type, adds up 
to thirty-five. Further, as far as the chromosomes are concerned, we 
cannot perceive any divergence of behaviour between the two. As in 
Rosa coriifolia, Fr. they appear in the heterotype division as seven bivalents 
and twenty-one univalents, and the descendants of these are distributed in 
the final microspores in the same general fashion. 
One important point of difference is manifested ; without exception, 
during synapsis and the stages intervening between that and diakinesis, 
the nucleolus appears to be doubled. Often enough, there are actually 
two nucleoli present ; less frequently they assume, an exact dumb-bell 
shape, whilst most commonly we have the appearance of a minor nucleolus 
attached to a greater. Whether this is to be interpreted as a vigorous 
budding of the nucleolus during the most critical period of the existence 
of the nucleus we know not, but we are not inclined to admit that 
explanation. 
In any case, the exact significance of the phenomenon, and why the 
two allied forms should so consistently differ in this respect, we cannot 
state. 
. (3) The Rosa glauca group. 
Rosa glauca will not delay us long. Just as in the other groups more 
than one microgene was examined, and these were Rosa Reuteri , God. 
(uniserrate), R. venosa , Schwartz, R. suberistata , Baker (biserrate), and 
R. stephanocarpa , Des. and Rip. (biserrate and with subfoliar glands). 
