1 82 Blackburn and Harrison. — The Status of the British 
when both are functional, the two gametes, the generative nucleus of the 
pollen grain and egg-cell of the same plant, carry widely different chromo- 
some numbers. Therefore reciprocal crosses between bushes of diverse 
chromosome complements should result in hybrids with different somatic 
counts, but when between members of the same gn^jup, as in the pentaploid 
^section, they should agree and both be pentaploid. In the only bush of the 
latter type examined the somatic number was thirty-five as demanded 
by theory. Moreover, as the bush grew adjacent to and resembled Rosa 
coriifolia var. Lintoni , it was a fair inference that that microgene had been 
the seed parent. In addition, since the other roses growing near by were 
all Rosa lutetiana , and in certain of its characters and ontogeny that form was 
approximated, Rosa lutetiana in all probability had acted as pollen parent. 
Now the egg-cell of Rosa coriifolia var. Lintoni would contain a nucleus 
with twenty-eight chromosomes, and' the generative nucleus of lutetiana 
pollen would probably possess seven, so that if these numbers have any 
value in determining the affinities of the hybrid it should resemble Lintoni 
more closely. Such was indeed the case, the lutetiana influence appearing 
mainly in the styles, stigmas, sepals, and young shoots, and less markedly 
in the prickles and glandular vestiture. 
Similarly, in the case of the Rosa Sabini , if Rosa sylvestris were the 
seed parent, then the hybrid should possess 144-28 (= 42) chromosomes ; 
if, on the other hand, it were the pollen parent, the hybrid number ought to 
be 14 + 7 = 21. Its somatic number was 42, whence we deduce that almost 
certainly the first conjecture was true, as was confirmed, firstly, by the fact 
that the bush in question grows in hedges amongst Rosa sylvestris and far 
from the nearest Rosa pimpinellifolia , and, secondly, it resembles Rosa 
sylvestris very much more closely than it does R. pimpinellifolia. 
The case of the Rosa hybrid pimpinellifolia x ( pimpinellifolia x coriifolia) ' 
is much more difficult to explain. If Rosa pimpinellifolia were the seed 
parent, as seems likely from the fact that the hybrid grows in a pimpinelli- 
folia colony on the Northumberland sand dunes, and fully 100 yards from 
the nearest coriifolia , the chromosome number should be 14 + 7 = 21 ; if the 
reverse held true, it ought to be 28 4- 14 = 42. Neither figure is correct, 
the somatic count being 28. 
Now an egg-cell of R. pimpinellifolia with fourteen chromosomes must 
form the starting-point, if we judge from the habitat of the plant. Suppose \ 
this to be fertilized by Rosa Coriifolia pollen with a nucleus endowed with 
seven chromosomes. The hybrid would thus display 14 + 7 = 21. In its 
megaspore formation, should events follow the usual course of Rosa 
hybrids, seven chromosomes would be shed, leaving an egg-cell with 
fourteen. As it is a matter of chance how the chromosomes are arranged in 
the partial reduction division, this fourteen would originate partly with 
coriifolia and partly with pimpinellifolia. Now for the pollination of this 
