78 
ANNE M. LUTZ 
nanella. The reverse combination should also reproduce the parental 
type. (2,) a b c d e f g -\- a' f g' = aa' bV cc' ee' ff' gg\ 0. La- 
marckiana. The first combination might be excused on the pretext 
of incompatibility, but this would hardly be sufficient to account for 
the absence of O. Lamarckiana from among the offspring of selfed 
nanella. Our difficulties are not lessened, as a little figuring will 
show, by assuming that two types of gametes are produced by one 
sex, and only one by the other, such as 9 abb — defg and a b c d e f g 
+ (J^ a' b'b'-d' e' f g' or 6" a' b' c' d' e' f g' ; or by assuming that 
all of the female gametes are of one type and all of the male of another 
type. Let us then consider the problem from another viewpoint. 
We may assume that these irregular 7-chromosome mutant gametes 
of 0. Lamarckiana, notwithstanding their numerous opportunities to 
unite with regular {Lamarckiana) 7-, are incapable of doing so, 
because of incompatibility, and that a gamete of this type can unite 
only with another of its kind: abb — defg-\-a' b'b' — d' e' j' g' = aa' 
bbb'b' — dd' ee' ff' gg', 0. nanella; also that the mutant produces 
male and female gametes of the same, single type. Our difficulties 
would still be with us, for 0. nanella X 0. Lamarckiana, and the 
reciprocal, would result in a new type (the same in both instances) 
quite unlike either parent. Furthermore, if the original irregular 
7-chromosome mutant gamete produced by 0. Lamarckiana were 
incapable of uniting with a regular y -Lamarckiana gamete in 0. 
Lamarckiana selfed, we would expect the two to be unable to unite 
in the crosses, yet we know that seed and offspring are readily secured 
from both. Even should we assume that the nanella group aa' bbb'b' 
— dd' ee' ff' gg', resulting from the union of identical gametes, produces 
female gametes of one type and male of another, such as a bbb—d—fg 
and a' b' — d' e'e' f g', our difficulties would not disappear. It seems 
impossible, on a chromosomal basis, to find an explanation for the 
fact that nanella and Lamarckiana, when selfed, produce only nanella 
in the first case, and only Lamarckiana in the second (barring rare 
exceptions), but that 0. nanella X 0. Lamarckiana and the reciprocal, 
produce both parental types in each case. Truly he who attempts 
to explain mutation on a strictly chromosomal basis finds his pathway 
beset with many obstacles. 
We do not know whether 15-chromosome mutant offspring of 14- 
chromosome 0. Lamarckiana result from unions of 9 7 with ^S, or 
9 8 with c^7, or from both combinations; since there is considerable 
