FIFTEEN- AND SIXTEEN- CHROMOSOME OENOTHERA MUTANTS IO9 
of chromosomes may be expected to breed true when the number is in 
excess of 14, except, perhaps, when this number is twice 14. Hence 
the fact that 0. semilata de Vries bred true, scarcely warrants the 
conclusion that this form probably had 16 chromosomes. We now 
know that 14-chromosome forms usually breed true and the evidence 
indicates that an occasional 15-chromosome form is also perfectly 
constant, while there are no records to show that offspring have been 
obtained as yet from forms known to have 16 chromosomes. It there- 
fore seems quite as probable that ^mi/a/a de Vries had 14, or even 15, 
as 16, chromosomes. 
10. Since somatic chromosome number has been shown to be 
constant in Oenothera, it is clear that unless 15-chromosome offspring 
are produced apogamously, or unless the chromosomes in excess of 15 
are eliminated after fertilization takes place, 15-chromosome forms can 
breed true, i. e., produce offspring having the somatic characters of 
the parent in every case, only when two gametes having dissimilar 
chromosome numbers, one odd and the other even, unite and produce 
viable seeds. While not all offspring resulting from such combinations 
reproduce the parental characters, it is certain that, with the excep- 
tions noted, they can be duplicated in no other way. 
11. The constancy or inconstancy of a plant is not determined 
solely by the presence of an even number of chromosomes in the first 
case and of an odd in the second. All depends upon the types of 
male and female germ-cells produced and the fertilization combinations 
which result in the production of seeds capable of germinating. 
12. As a rule, larger quantities of seeds are obtained from 14- 
chromosome forms selfed, or pollinated by other 14-chromosome 
forms of the same, or different species, than from 14 +-chromosome 
forms selfed, or pollinated by other 14 -f -chromosome plants of the 
same, or different species, particularly if the 14 -chromosome 
individuals have more than 14, but fewer than 28, chromosomes; 
furthermore, higher percentages of germination are usually secured 
from the former than from the latter when seeds not more than one 
year old are sown in pans of sterilized soil in January and are kept 
under ordinary greenhouse conditions. 
13. The number of seeds produced by a form and the ability of the 
seeds to germinate, at least within the time limits specified, are factors 
which appear to be associated with the chromosome number of the 
plant, or numbers of the plant, producing them. 
