1 68 Blackburn and Harrison. — The Status of the British 
a horseshoe shape with some of the univalents included' within its arms, 
although most are arranged so as to surround them partially. Both at 
earlier stages and subsequently matters move in the very strange fashion 
described at some length for R. coriifolia and, save that the bivalents in 
this case are seven in number, for R. Sahini. 
In the interkinesis we see two major nuclei, with varying numbers of 
micronuclei as well as with odd degenerating chromosomes. In the 
homotype the movement of the chromosomes to the spindle is even more 
disturbed by the hybrid combination at work, and despite the fact that the i 
scheme followed is that typical of pentaploid forms, the tetrad is so very s 
abnormal that almost the whole of the microspores are abortive. Two 
per cent, of good grains was the best count made from a microscopical 
inspection of the contents of a number of anthers. 
Similarly, megaspore formation cannot have been uniformly successful, 
for, although the bush produces some well-developed seeds by a process 
of apomixis, only a small proportion of the fruits swell satisfactorily as the 
seeds ripen within ; the remainder simply contain a mass of chaffy scales. 
IV. The Cytology of the Pentaploid Group. 
(a) The Afzelianae. 
The section Afzelianae comprises all of the segregates classified by 
English botanists, as genuine species, under the old collective names Rosa \ 
coriifolia and R. glauca However, so closely do the forms assigned to one j 
or other of these supposed cardinal species approach one another, that no 
difficulty is encountered in finding two bushes absolutely identical, save for 
some microscopic pilosity on the midribs of the leaflets of one, which 
suffices to drive that form into R. coriifolia and the other into R. glauca. 
From this the recognition of the two seems ridiculous, and to avoid their 
unnatural separation Almqvist united them under the oldest collective name ! 
capable of covering both sets of plants ; that name was Rosa Afzeliana. 
At the same time, he removes the glandular microgenes, considered by 
most workers on the genus Rosa to be true segregates of R. coriifolia , 
to the section Rubiginosae. In this, whilst we recognize some sound 
reasons for the procedure, we are unable to follow him ; the gradual 
transition from R. coriifolia (type) to R. Lintoni renders it almost 
impossible to differentiate virescent forms near the type from the less 
glandular varieties of the roses so transferred. Thus, our conception 
of the Afzelianeae differs from what Almqvist includes under the 
name, inasmuch as, in our opinion, the section contains the whole of the 
old ‘ species ’ R. coriifolia and R. glauca. 
(i) R. coriifolia , Fries. 
This plant was almost the first to which we turned our attention. We 
were very fortunate in our choice. In it we secured almost ideal material 
