parents, suggesting that Tripsacum germplasm produces 
effects in this direction. There is also reason to expect 
number of knobs to be associated with length of ear, for 
the C segment from teosinte increases the length of the 
ear in segregates from crosses of maize and teosinte. How- 
ever, this occurs only when the C segment is acting alone. 
When other segments are present the ears are shorter. 
No significant difference between low and high groups 
is shown in average length of ear on the basis of a four- 
fold table. There are some extremely long ears in the 
collection, however, and these are confined to varieties 
with intermediate knob numbers. None is found in the 
group with very low or very high knob numbers. 
The groups do not differ significantly in diameter at 
base of ear and the difference in diameter at tip of ear is 
barely significant. here is a highly significant difference 
in the tip-base index which is derived by dividing diame- 
ter at tip by diameter at base. This tendency for the low- 
knob group to have tapering ears is shown still more 
strikingly in the taper index which expresses as a per- 
centage the amount of reduction in diameter which oc- 
curs in each centimeter unit of length. ‘Taper index is 
derived by dividing by length, the percentage difference 
between diameters at base and tip. The relationship be- 
tween knob number and taper index is approximately 
linear. 
Number and regularity of the rows 
Another effect which Tripsacum germplasm produces 
in hybrids of maize and teosinte is a reduction in the 
number of rows of grain and a straightening of the rows. 
We might therefore expect both of these characters to 
be associated with knob numbers. The second expecta- 
tion has been met but the first has not. The difference 
in the average number of rows in the low-knob and high- 
[ 227 ] 
