numerous and are terminal; in maize and Mexican teo- 
sintes they are largely internal. The differences which 
now exist are assumed to be the result, not of structural 
changes, but of a few mutations which have affected 
chromosome gradients in such a way that knob positions 
in maize and Mexican teosintes are, on the average, closer 
to the centromere than in Guatemalan teosintes. Knob 
forming points, even though not visible, are regarded as 
always present. Knobs are formed only when adequate 
amounts of pycnotic material are available. Abundance 
of pycnotic material is associated with primitiveness. 
Tripsacum is regarded as the most primitive of the Amer- 
ican Maydeae, standing near the base of the tree, teo- 
sinte from eastern Guatemala is regarded as the trunk of 
the tree; teosinte from western Guatemala as a major 
branch; maize and Mexican teosintes as the finer branch- 
es. It is suggested that maize originated directly from 
one of many teosinte strains. However, the suggestion, 
(Harshberger, 8; Collins, 5) that it has been derived from 
a hybrid of teosinte and an unknown grass, is not ruled 
out. 
If we consider only the conclusion that an abundance 
of pycnotic material is associated with primitiveness, the 
relationship between knob numbers and other character- 
istics agrees almost as well with Longley’s hypothesis as 
with that of Mangelsdorf and Reeves. In either case re- 
lationships between knob numbers and other characteris- 
tics are to be expected. The chief difference is that in one 
‘ase the association is with ‘‘primitive’’ characteristics, in 
the other with ‘‘Tripsacoid’’ characteristics. The terms 
are partially synonymous but not completely so. Long- 
ley’s hypothesis falls short in so many additional respects, 
however, that an interpretation of our data in terms of 
his hypothesis is virtually precluded. 
Longley’s conclusions are based almost wholly upon 
[ 240 | 
