Feb. 23,1924 
Teosinte Maize Hybrids 
561 
III. In attempting to determine to what extent the crinkly complex of 
characters tends to remain in combination it is desirable to eliminate as 
far as possible the correlations expected on physiological grounds. Thus 
there are logical and experimental reasons for expecting physiological 
correlations of height with the five following characters given in Table 
III: Central spike, branching space, number of branches, length of 
longest branch, and length of blade. The nature of the crinkly teosinte 
cross is such as also to provide genetic causes of correlations of height 
with these characters, but with width of leaf genetic factors presumably 
operate in one direction while physiological factors operate in the other. 
Thus, the crinkly variation has short stalks and wide leaves as com¬ 
pared with teosinte, and if there is a genetic correlation between these 
characters a negative correlation between height and width of leaves 
would be expected in the F 2 . Further, since physiological factors nor¬ 
mally tend to produce a positive correlation between height and width 
of leaves, the net effect of genetic and physiological factors operating in 
opposite directions would be to reduce the correlation essentially to 
zero. This is the observed result, though the coefficient is negative 
(r= —0.041), indicating possibly that the genetic factors were somewhat 
more effective than the physiological factors in bringing about the ob¬ 
served relationship. 
When the correlation between these two characters of the Fj plants 
is examined it also is found to be negative. While the correlation is 
not significant, its departure from the 0.2 or 0.3 usually obtained with 
these characters and the fact that the correlation with leaf index also is 
negative seem to require some explanation. It can be assumed that some 
peculiar environmental condition caused an increase in height and an 
actual and relative decrease in width of leaf such as obtain under extreme 
conditions where the plants become spindling; but such an hypothesis 
can not be accepted with favor when the exceptionally sturdy nature of 
the plants is taken into account. 
The alternative is to attribute the results to genetic causes. Thus, if 
either the crinkly or the teosinte parent were heterozygous for a dominant 
factor which controlled a character complex of height, length, and width 
of blade, and the corollary height factors, length of central spike and 
branching space, then the F x in effect would be a back cross of this nature 
and the observed F 1 correlations would be obtained. While such a con¬ 
dition is not beyond the realm of possibility, it certainly seems singular 
that the complex of characters involved is so essentially that by which 
the recessive variation crinkly differs from the normal form. 
It may be less unreasonable to assume that in hybrids such as this, 
where a general blending of the characters takes place, occasionally in 
certain plants one parental type or the other becomes partially dominant. 
There then would be F t individuals in which all the maize characters 
would be raised somewhat above the mean, and others in which they 
would be below the mean. Such behavior would result in a population 
consisting largely of plants intermediate between the parents, but with 
the tails of the distribution made up of the parental character combi¬ 
nations, being those cases where the threshold of dominance had been 
crossed. The second generation of these several F x plants need not 
necessarily behave very differently, and the F t correlations would be 
provided for. 
It is well known that the degree of dominance is extremely variable, 
but this variability formerly has been thought of in connection with indi¬ 
vidual characters. There is evidence from other sources that variations 
