known genetic and cytological mechanisms, the assump- 
tion that the knob positions in maize have been derived 
from those in teosinte, not by structural changes in the 
chromosomes, but as the result of a few mutations affect- 
ing the gradient. There is no doubt that chromosome 
behavior is partially gene-controlled. There are at least 
five genes in maize: asynaptic, sticky, polymitotic, and 
variable-sterile (two genes) affecting chromosome behav- 
ior and cell division (cf. Rhoades & McClintock, 20). 
In Allium there is a situation with respect to chiasmata 
which is almost identical with that of maize and teosinte 
with respect to knobs. In A. fistulosuwm the chiasmata 
are interstitially localized; in A. cepa the chiasmata are 
all terminal. The difference is due to a single recessive 
gene (Iimsweller, 7; Levan, 12). But in this case as well 
as in similar cases all chromosomes in a single nucleus 
follow the same pattern; all are controlled by the same 
nuclear mechanism. In the Fy hybrid of A. fistulosumX 
A. cepa all chiasmata are terminal. If the knob gradients 
in maize and teosinte are analagous, one must expect a 
decided shift in knob positions in Fy hybrids of maize and 
teosinte. The internal maize knobs should move to the 
chromosome ends, most of the teosinte knobs should 
move to internal positions or both should move to inter- 
vening regions depending upon whether the mutant 
maize-gradient-controlling genes are recessive, dominant 
or intermediate in their effects. Such wholesale shifts 
have not been reported. In fact one has but to examine 
Longley’s (13) illustration of the chromosomes in the 
KF, hybrid of maize and Florida teosinte to be convinced 
that they do not occur. 
Additional assumptions can be made, assumptions in- 
volving large numbers of mutant genes and the differ- 
ential control of maize-gradient and teosinte-gradient 
chromosomes within the same nucleus. Longley (15) 
[ 242 ] 
