550 Gates.—The Trisomie Mutations of Oenothera. 
that it must have the same unbalanced chromosome condition. 1 The genetic 
behaviour of all these forms will be referred to later. The mutant incurvata 
(see Table I) should probably not be included in this series, for, while it is 
evidently a distinct type, it appeared in the Swedish race of Oe. Lamar ck- 
iana and may owe its peculiarities to certain genetic differences which exist 
between it and the Lamarckiana of de Vries’s cultures. 
We now come to the various 15-chromosome forms described by Miss 
Lutz ( 1917 ). In the first place is subovata , de Vries, a sterile form which 
de Vries ( 1909 ) obtained not only from Lamarckiana and lata x Lamarck¬ 
iana , but also from sublinearis and from scintillans. The latter has 15 
chromosomes (Hance, 1918 ), and the genetic behaviour of sublinearis indi¬ 
cates that it also probably had 15. Elliptica , de Vries, also, from its genetic 
behaviour, is probably in the same condition. Miss Lutz describes a 15- 
chromosome form which she calls bipartita and which is much more like 
Lamarckiana than most of these types. It differs in that flowers with more 
than four stigma-lobes are much commoner than in Lamarckiana , and the 
petals frequently develop diagonal clefts owing to their not overlapping in 
regular order. Cleft petals, owing to conditions which are apparently 
somewhat different, have been observed in various other forms (van 
Overeem, 1920 ; Gates, 1917 , 1923 ). The flowers are smaller and the leaves 
more finely crinkled than in Lamarckiana. Much of the pollen is obviously 
bad and it was very difficult to get seeds from self-pollination. There is 
apparently no genetic evidence to indicate the relation of bipartita to other 
15-chromosome mutants, except that in one instance Lamarckiana x lata 
gave two bipartita (?) in addition to 63 Lamarckiana , one lata , and 
certain other forms. Miss Lutz also refers to two other mutations from 
Lamarckiana with 15 chromosomes, but they are not described. They are 
‘ type 5509 ’, which is said to be a modified oblonga , and ‘ type 2806 ’, which 
is said to have ‘ many points in common with type 5509 \ Both then appear 
to be nearest oblonga. In addition to the above 15-chromosome mutations 
from Lamarckiana , others have so far been obtained only from lata selfed or 
from lata x Lamarckiana. These will be referred to later. 
For the purpose of further analysis it appears best to confine ourselves 
largely to the eight and more forms with an extra chromosome which have 
been identified by de Vries. The genetic relationships of these forms to 
one another renders the whole problem a very difficult one. The next 
evidence to be considered concerns the genetic behaviour of these forms 
when selfed or in crosses. In general they are more or less completely 
male sterile, and when pollinated from Lamarckiana give the two parental 
types with a greater or lesser preponderance of Lamarckiana offspring. 
Eight forms arising directly from the Lamarckiana of de Vries’s cultures and 
1 Van Overeem (1922) has recently shown that the somatic cells of cana have 15 chromosomes. 
