i8o Blackburn and Harrison . — The Status of the British 
for that. But in all of their irregular behaviour at this stage — the failure of 
complete pairing, the lagging of chromosomes both in passing to and from 
the equatorial plate, the formation of micronuclei, the accentuation off these 
features in the homotype division, and, finally, the generation of defective 
pollen — the known rose hybrids differ in no way from what we have 
pictured for us by Rosenberg as occurring in his Droserae, Tischler in his 
Bryoniae, and by ourselves, amongst others, in animal hybrids. We must, 
therefore, assign the same irregular processes encountered at the same 
stages in Rosa to the same cause, and that cause hybridity. Once having 
admitted this in the case of the patent rose hybrids, to the same disturbing 
influence we must look for our explanation of precisely the same unusual 
occurrences in the other forms ; they are assuredly latent hybrids. In this 
manner, at one stroke, by invoking hybridity we account for the extraordinary 
phenomenon of pentaploid species, the harmony between their cytology and 
that of the tetraploid Villosae, and lastly, the whole chain of circumstances 
arising from the peculiarities in their meiosis. 
Digby, following Juel, has urged that all of the above-mentioned 
irregularities, beginning with the untidy spindle figures and culminating 
with abnormalities of chromatin distribution in the heterotype and the homo- 
type divisions, cannot be regarded as conclusive proofs of hybridity, and 
instances the classic case of Hemerocallis fulva as proving that such 
behaviour occurs in pure species. To us this seems to savour strongly 
of begging the question, for not a vestige of proof has ever been adduced to 
substantiate the claims of Hemerocallis , and other plants exhibiting similar 
divergences from typical microspore formation, to be regarded as ‘ pure 
breeds \ On the contrary, if we might be permitted to base our opinions on 
the definite proofs provided by absolutely the same phenomena detected 
by Rosenberg in artificial hybrids in Hieracium and by ourselves in the 
patent crosses in Rosa , and furthermore by aiding our judgement by evidence 
derived from other instructive similarities, we feel sure that in the case of 
the plant treated as the crucial example we have a persistent, if more or 
less latent, hybrid like our roses. 
Comparing the pseudo-reduction in the pentaploid roses and their 
tetraploid associates, and backing up the comparison with what obtains in 
Rosa Sabiniy we are bound to admit that only one satisfactory explanation 
exists for their common peculiarities. Just as the ten chromosomes derived 
from Drosera rotmidifolia find mates in the reduction division of D. obovata 
in ten provided by the D. longifolia parent, so one must conclude that the 
common seven running through the whole Caninae supersection has been 
introduced by some species (quite possibly Rosa arvensis ) possessing that 
number as its haploid complement, when it has crossed with a microgene or 
microgenes with a haploid number of twenty-one in the Villosae and 
twenty- eight in the rest of the Caninae. This notion might seem to demand 
