Gates .— The Trisomic Mutations of Oenothera. 551 
showing this type of behaviour are known to have 15 chromosomes. Miss 
Lutz has shown that subovata also has 15 chromosomes, and the same may 
reasonably be inferred of sublinearis and clliptica from their genetic 
behaviour. In addition, we have bipartita , making a total of twelve forms* 
‘Type 5509 1 and ‘ type 2806 ’ of Miss Lutz may be looked upon as modi¬ 
fied oblongas, the modification being due, perhaps, to the presence of inde¬ 
pendent genetic factors, although in Oenothera so few factors of this kind 
have been demonstrated that great weight cannot be attached to this sug¬ 
gestion. 
We may next consider the relationships of the trisomic forms as indicated 
by the offspring they produce when selfed or pollinated from Oe. Lamarck- 
iana . As formerly pointed out (Gates and Thomas, 1914 ), lata and semi - 
lata , Gates, are peculiarly related, since the latter is essentially intermediate 
between lata and Larnarckiana and is only known to occur in the offspring 
of lata x Larnarckiana. Similarly mut. superfiua , de Vries ( 1916 ), occurred 
as a mutant from lata (presumably pollinated by Larnarckiana) in 1914, and 
when selfed gave the usual dimorphic progeny characteristic of trisomic 
forms. It probably had 15 chromosomes. Then cana , pallescens , Lactuca , 
and liquida are found by de Vries ( 1916 ) to arise apparently with equal 
facility from Larnarckiana or from lata x Larnarckiana. Indeed, with far 
fewer seeds developed, owing to the sterility, they appear to arise with 
a much higher frequency from the latter source. Mut. subovata has also 
arisen both from Larnarckiana and from lata x Larnarckiana. Furthermore, 
the ‘ type 4499 ’ of Miss Lutz has arisen from lata x Larnarckiana and from 
lata self-pollinated, while her cxilis , exundans , and ‘ type 5365 * are only 
known from the offspring of Oe. lata selfed. Hence there appears to be 
a peculiar relationship between lata and some at least of the other trisomic 
forms. 
Oc. cana mutants have been obtained (de Vries, 1916 ) from Larnarckiana , 
lata , pallescens , scintillans , and laevifolia , but its frequency as a mutant from 
Larnarckiana is only 0-03 per cent., while from lata it may be as high as 
9 per cent. When selfed, cana gives, in addition to cana and Larnarckiana, 
a significant frequency of nanella , as well as occasional stray mutations, in¬ 
cluding albida. Pallescens , which is said by de Vries to differ the least 
from Larnarckiana , was not discovered until 1911. It has since been 
obtained from some of the derivatives of Larnarckiana. When selfed it 
gives, in addition to the usual dimorphic progeny, cana , liquida , scintillans , 
lata , albida , and rubrinervis mutants, the mutants reaching as high as 4 per 
cent. 1 Mut. Lactuca , first observed in the offspring of lata x Larnarckiana 
in 1913, gave when selfed 39 Larnarckiana , four Lactuca , and one 
nanella. Mut liquida most resembles scintillans. It was first observed in 
1 The question ol rubrinervis will be discussed below. 
