400 LOTSY, CURRENT THEORIES OF EVOLUTION. 
improbable that pollengrains with more than 7 chromosomes do 
not sprout, because v. OVEREEM found in the case of triploid Oeno- 
thera’s that, although pollen with all chromosomenumbers between 
7 and 14 was formed, only such as contained either 7 or 14 chro- 
mosomes was able to fertilise the eggcells. | 
Selffertilisation however is rare in the Rosae; in most cases they 
reproduce apomictally — possibly from nucellar buds, how is not 
yet clear — which of course contain, as all vegetative tissue 35 
chromosomes, so that the greater part of the Rosae of Europe 
are very ancient F, hybrids, arisen thousands of years ago, and 
conserved as such by apomictal reproduction. A certain amount of 
diversity of them can of course be caused by selffertilisation as » 
the 7 disomes allow mendelian segregation. TäCKkHOLM himself 
assumes for part of the diversity vegetative mutation. He says: 
„The enormous polymorphy characterising the section Caninae 
might, to-a large extent, be ascribed to this obvious hybridism. 
However mutation phenomena have no doubt contributed to the 
polymorphy. In this case, the mutations must be vegetative, pro- 
duced in apomictal stocks. Matsson (Svensk Bot. Tidskr. 1912) 
has in a sowings of tomentosa* subcristata found some forms dif- 
fering from the mother. The flowers had not been castrated. 
Nevertheless, that one of these forms which I examined, probably 
was not produced in a sexual way, is evident from the chromo- 
someset being quite the same as that of the investigated tomentosa- 
forms representing the F;-type (7 gemini and 21 unpaired chro- 
mosomes). 
Of course the possibility of mutation exists — as remarked be- 
fore in heterozygous forms also” — but the evidence for it must 
be very much better founded than the one case, here mentioned: 
a deviating form obtained from a not castrated heterozygous plant, 
which plant certainly produced pollengrains with 7 chromosomes, 
(even if it produced viable ones with a higher number) so that 
the cytological evidence does not at all exclude the possibility of 
selffertilisation. 
Hybrids between a Rosa canina with 28 chromosomes and 
another species with 7 chromosomes in its pollen, can of course, 
if selffertilised, segregate as far as the chromosomes which enter 
into pairs with one another are concerned, but will possess all 

5 
