176 Blackburn and Harrison . — The Status of the British 
all an extreme example of this, and b shows the division into the usual octad 
in which almost all of the pollen grains are multinucleate (see also 
PI. X, Fig. 43 ). 
This accentuated irregularity would seem to suggest that some of the 
tomentosa forms may be very recent, nay even F 2 , crosses. PI. X, Fig. 45 
illustrates the regularity of the somatic mitosis even here. 
V. The Cytology of the Tetraploid Types. 
The Villosae. 
This type, as far as the British roses go, comprises within its limits 
nothing but Villosa. segregates, and amongst those prepared for study were 
representatives of the Rosa mollis group (R. mollis , Sm., and R. coerulea, 
Woods) as well as of the R. omissa 1 group (R. omissa , Deseg., R. suberecta , 
Ley., and R. Sherardi , Dav.). 
In the case of four of these no further indications of the plant intended 
are necessary, but, in view of Tackholm’s 2 pronouncement as to the uncon- 
formability of R. omissa to the tetraploid type, and our proof that it 
is tetraploid, the exact plant we studied must be defined with complete 
accuracy. 
Probably no British worker has a better comprehension of R. omissa 
than the veteran W. Barclay, and direct comparisons made between his 
gatherings and Swiss type-forms from Ddseglise (the describer of the species) 
have demonstrated beyond cavil that they must be regarded as conspecific. 
We possessed specimens of Barclay’s collecting and naming, and with 
these the bushes whence we derived our material are practically identical. 
Further, we are in a position to confirm the accuracy of our determination 
indirectly. As Almqvist, the well-known Swedish rhodologist who named 
Tackholm’s plants, needed British roses for examination in order to fix the 
range of the forms of his new system (and not because we doubted our own 
namings), we sent him a score of Northumberland and Durham examples with 
three of Barclay’s; all, without exception, were returned as Rosa molli- 
trachyphylla , Almq. This confirms our belief that Barclay and ourselves 
were dealing with the same plant and also, as we already knew, that our 
omissa was a Villosa rather than the Tomentosa form it had been assumed 
to be. What Almqvist and therefore Tackholm understand by Rosa 
omissa will appear as a sequel in our concluding remarks and thus reveal 
the reason for Tackholm’s curious findings. 
The meiotic phase in these forms is uniform for all, and alfowing for 
its harmony with the same period in R. Sabini and R. coriifolia one brief 
description will suffice. It is ushered in by a typical synapsis, opening out 
1 Also known sometimes as the R. Sherardi or even the R. subglobosa group. 
2 This paragraph and the next were inserted after we had received a copy of Tackholm’s brief 
paper* 
